Prediction method and device of destination location, computer device and storage medium

Abstract

The present application relates to the technical field of vehicle data prediction, and discloses a destination position prediction method and device, computer equipment and a storage medium, the method comprising: obtaining a departure position and a departure time of a vehicle; obtaining a first grid corresponding to the departure position in a target area, and querying a first candidate trip matching the departure time based on a historical trip set involving the first grid and adjacent grids corresponding to the first grid; performing correlation analysis based on the first candidate trip and a real-time position of the vehicle to obtain a target candidate trip of the vehicle; and performing median center calculation using a trip end point in the target candidate trip to obtain a predicted destination of the vehicle. The present application introduces real-time position information of the vehicle for correlation analysis, and corrects the candidate trip set as the trip changes, thereby ensuring that the latest position data obtained after the trip starts can be effectively utilized to gradually approach the real destination of the vehicle, and the accuracy and timeliness of destination prediction are improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle data prediction technology, and more specifically to a method, apparatus, computer equipment, and storage medium for predicting destination location. Background Technology

[0002] To improve the overall efficiency of hybrid vehicles, the industry widely researches destination prediction strategies based on navigation information. These strategies primarily optimize the operating limits of the engine and electric motor by obtaining information about the destination and road conditions ahead in advance, thereby improving energy efficiency. Furthermore, accurate destination prediction is also a key technological foundation for achieving location-based precision services and smart cities and intelligent transportation systems.

[0003] However, navigation is rarely used on routes familiar to drivers, making it difficult for the system to reliably obtain destination information and thus preventing optimization strategies from being implemented. Although related technologies attempt to address this issue by using historical data for destination prediction, they generally suffer from insufficient data utilization, particularly failing to effectively utilize the latest information acquired in real time as the vehicle travels after the trip begins, resulting in limited prediction accuracy. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method, apparatus, computer device, and storage medium for predicting destination location, in order to solve the problem of how to effectively utilize the latest location information acquired in real time as the vehicle travels after the start of the trip, thereby improving the accuracy of destination prediction.

[0005] In a first aspect, embodiments of the present invention provide a method for predicting the destination location, which obtains the vehicle's departure location and departure time; Obtain the first grid corresponding to the departure location in the target area, and query the first candidate trip that matches the departure time based on the historical trip set involved in the first grid and the adjacent grids corresponding to the first grid. The target area is the geographical area covered by the vehicle's travel, and the target area includes multiple uniformly divided grids. Based on the correlation analysis between the first candidate route and the real-time location of the vehicle, a target candidate route with respect to the vehicle is obtained; The predicted destination of the vehicle is obtained by calculating the median center using the endpoints of the target candidate routes.

[0006] This invention introduces real-time vehicle location information for correlation analysis and corrects the candidate trip set as the trip changes, ensuring that the latest location data obtained after the trip begins can be effectively utilized, thereby gradually approaching the vehicle's true destination and improving the accuracy and timeliness of destination prediction.

[0007] In conjunction with the first aspect, in one implementation, obtaining the first grid corresponding to the starting position in the target area includes: Obtain the vertex position of any grid vertex in the target region; Calculate the lateral and longitudinal distances of the starting position relative to the vertex position; Using the lateral distance and the longitudinal distance, the first grid into which the starting position falls is determined.

[0008] This invention maps continuous, unstructured vehicle location data to structured grid cells by calculating the horizontal and vertical coordinate distances between the starting point and preset grid vertices, laying a reliable data foundation for subsequent grid-based historical trip matching and effectively improving the initial accuracy of destination prediction.

[0009] In conjunction with the first aspect, in one implementation, the step of performing real-time correlation analysis based on the first candidate trip and the real-time location of the vehicle to obtain a target candidate trip related to the vehicle includes: Query the first candidate itinerary for a second candidate itinerary associated with the departure location; Obtain the real-time location of the vehicle; Select target candidate routes from the second candidate routes that pass through a set range around the real-time location.

[0010] This invention first screens out second candidate routes based on the departure location, and then uses real-time location to perform precise spatial correlation analysis, retaining only those routes whose historical trajectories highly match the vehicle's current actual travel path. This ensures that the latest location information obtained after the start of the trip can be effectively utilized to gradually narrow down the prediction range, thereby improving the accuracy and real-time performance of destination prediction.

[0011] In conjunction with the first aspect or its corresponding implementation, in one implementation, querying the first candidate itinerary for a second candidate itinerary associated with the departure location includes: Obtain the starting point of the first trip in the first candidate trip; Calculate the actual distance between the starting point of the first journey and the starting position; The first candidate route whose actual distance is less than a preset distance threshold is determined as the second candidate route.

[0012] This invention calculates the actual distance between the starting point and the destination and performs threshold filtering to obtain historical routes that are strongly spatially correlated with the current trip, while excluding historical trips with excessively long destinations or that are obviously irrelevant. This optimizes the quality of candidate trips before using real-time location for fine filtering, laying a reliable data foundation for subsequent accurate prediction.

[0013] In conjunction with the first aspect or its corresponding implementation, in one implementation, the step of filtering target candidate trips from the second candidate trips that pass through a predetermined range around the real-time location includes: Obtain the second grid corresponding to the real-time location; Obtain the adjacent grids corresponding to the second grid from the target region; From the second candidate routes, query the third candidate routes that pass through the second grid and the corresponding adjacent grids of the second grid; The set range is determined based on the real-time location as the center and the preset distance threshold as the radius; The third candidate route that falls within the set range is determined as the target candidate route.

[0014] This invention first uses the grid where the real-time location is located and its neighborhood to perform a fast search and narrow down the candidate range; then it makes a final judgment based on the precise geometric neighborhood of the real-time location, which takes into account both computational efficiency and spatial matching accuracy, and ensures that the latest travel trajectory of the vehicle can be dynamically captured, thereby filtering out the historical journeys most relevant to the current path.

[0015] In conjunction with the first aspect, in one implementation, the step of calculating the median center using the destinations of the target candidate trips to obtain the predicted destination of the vehicle includes: Calculate the arithmetic mean center of the trip endpoints, and calculate the weighting factor for each trip endpoint based on the arithmetic mean center; The predicted destination of the vehicle is obtained by weighting the trip endpoint according to the weighting factor.

[0016] This invention achieves more accurate and robust destination prediction by assigning weight factors to different trip endpoints and performing weighted calculations, ensuring that the final predicted destinations tend to cluster in the high-frequency core area of ​​historical trip endpoints, thereby effectively improving the accuracy and reliability of the prediction results.

[0017] In conjunction with the first aspect, in one implementation, after calculating the median center using the destinations in the target candidate trips to obtain the predicted destination of the vehicle, the method further includes: Detect whether the real-time location of the vehicle has been updated; If the real-time position is updated, detect whether the second grid corresponding to the updated real-time position has changed; If the second grid changes, the destination prediction steps are re-executed based on the changed second grid; or, if the second grid does not change, the current destination prediction result is maintained until a preset duration after which the prediction is terminated.

[0018] This invention continuously monitors vehicle position and grid changes to capture deviations in the vehicle's trajectory. When the trajectory changes significantly, the destination is re-predicted, ensuring the prediction results are always synchronized with the latest driving status. It not only fully utilizes real-time location information throughout the journey to continuously correct predictions but also avoids unnecessary computational overhead through a delayed termination mechanism when the grid remains unchanged. This optimizes resource utilization while ensuring prediction accuracy and real-time performance.

[0019] Secondly, embodiments of the present invention provide a destination location prediction device, the device comprising: The data acquisition module is used to obtain the vehicle's departure location and departure time; The trip query module is used to obtain the first grid corresponding to the departure location in the target area, and query the first candidate trip that matches the departure time based on the historical trip set involved in the first grid and the adjacent grids corresponding to the first grid. The target area is the geographical area covered by the vehicle's travel, and the target area includes multiple evenly divided grids. The trip analysis module is used to perform correlation analysis between the first candidate trip and the real-time location of the vehicle to obtain the target candidate trip related to the vehicle. The location prediction module is used to calculate the median center using the destinations in the target candidate routes to obtain the predicted destination of the vehicle.

[0020] Thirdly, embodiments of the present invention provide a computer device, including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the destination location prediction method of the first aspect or any corresponding embodiment described above.

[0021] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions for causing a computer to execute the destination location prediction method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0022] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1This is a flowchart illustrating a method for predicting the location of a destination according to some embodiments of the present invention; Figure 2 This is a grid diagram of the target area in a destination location prediction method according to some embodiments of the present invention; Figure 3 This is a structural block diagram of a destination location prediction device according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of 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, 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.

[0025] According to an embodiment of the present invention, a method for predicting the location of a destination is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] This embodiment provides a method for predicting destination location, which can be used in vehicle-mounted terminals. Figure 1 This is a flowchart of a destination location prediction method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain the vehicle's departure location and departure time.

[0027] Step S102: Obtain the first grid corresponding to the departure location in the target area, and query the first candidate trip that matches the departure time based on the historical trip set involved in the first grid and the adjacent grids corresponding to the first grid. The target area is the geographical area covered by the vehicle's travel, and the target area includes multiple uniformly divided grids.

[0028] Step S103: Based on the correlation analysis between the first candidate route and the real-time location of the vehicle, the target candidate route with respect to the vehicle is obtained.

[0029] Step S104: Calculate the median center using the destinations in the target candidate routes to obtain the predicted destination of the vehicle.

[0030] The destination location prediction method provided in this embodiment obtains the vehicle's departure location and departure time; obtains the first grid corresponding to the departure location in the target area, and queries the first candidate route that matches the departure time based on the historical trip set involved in the first grid and its adjacent grids; performs correlation analysis between the first candidate route and the vehicle's real-time location to obtain the target candidate route; and calculates the median center using the destination of the target candidate route to obtain the vehicle's predicted destination. This embodiment introduces the vehicle's real-time location information for correlation analysis and corrects the candidate route set as the trip changes, ensuring that the latest location data obtained after the trip starts can be effectively utilized, thereby gradually approaching the vehicle's true destination and improving the accuracy and timeliness of destination prediction.

[0031] For step S101, obtain the vehicle's departure location and departure time.

[0032] In one embodiment, the vehicle's departure location and departure time at the start of the current trip are first obtained. The departure location can be composed of latitude and longitude coordinates provided by the vehicle's onboard Global Positioning System (GPS) or a positioning module in other vehicles, while the departure time is the timestamp of the trip's departure moment.

[0033] For step S102, the first grid corresponding to the departure location in the target area is obtained, and the first candidate trip that matches the departure time is queried based on the historical trip set involved in the first grid and the adjacent grids corresponding to the first grid. The target area is the geographical area covered by the vehicle's travel, and the target area includes multiple uniformly divided grids.

[0034] In one embodiment, to manage and spatially query unordered vehicle travel trajectories in a normalized manner, a target area is predefined. The target area refers to the geographical range covered by vehicle travel activities, such as a complete urban administrative region. Within this target area, a travel database containing all private car historical travel data is constructed. The travel database records key information for each private car travel trajectory, such as the origin, destination, waypoint sequence, and departure time.

[0035] Reference Figure 2This diagram illustrates a grid of the target area in a destination location prediction method according to some embodiments of the present invention. Specifically, the entire target area is uniformly divided into a grid array consisting of N rows and N columns according to a preset size, thus forming N×N regularly arranged grids, each of which constitutes a basic geographic indexing unit. The side length of a single grid needs to be determined through vehicle calibration. The vehicle calibration process is affected by vehicle density and regional characteristics to ensure that the grid size achieves a balance between spatial accuracy and computational efficiency, such as the difference in vehicle density between city centers and suburbs, road network structure, and functional zone division. For example, a smaller grid size can be used in densely populated city centers to obtain higher accuracy, while a larger grid size can be used in sparsely populated suburbs to reduce computation. Each grid is assigned a unique identifier. Preferably, a combination of row and column numbers is used for location encoding, simplifying the grid positioning and neighborhood search logic. Furthermore, since the grids of the target area in this embodiment are numbered starting from

[00] , the actual row and column corresponding to each grid number are larger than the corresponding numbers in the number. For example, the grid coded as

[35] actually represents the grid in the 4th row and 6th column.

[0036] In one embodiment, the vertex position of any predefined grid vertex in the target area is obtained. For example, the origin of the entire grid array can be selected, i.e., the top-left vertex of the grid in row 0 and column 0, which has known latitude and longitude coordinates. Next, the lateral and longitudinal distances of the vehicle's starting position relative to this vertex position are calculated, i.e., the offset of the vehicle's starting position in the latitude and longitude directions. Finally, using these two distance values, combined with the known grid size, the specific grid in which the starting position falls can be uniquely determined; this grid is defined here as the first grid. Specifically, the number of each grid is determined by the following formula: Column number = floor((horizontal distance) / horizontal grid size); Row number = floor((vertical distance) / vertical grid); Position code C = (row number, column number) For example, if the grid size is 10KM*10KM, and the horizontal distance from the calculated trajectory point to the upper left corner

[00] is 14km and the vertical distance is 27km, then its grid number is

[21] . Specifically, the steps are as follows: Get the vertex position of any grid vertex in the target area; Calculate the lateral and longitudinal distances of the starting position relative to the vertex position; Using the horizontal and vertical distances, determine the first grid into which the starting position falls.

[0037] This embodiment maps continuous, unstructured vehicle location data to structured grid cells by calculating the horizontal and vertical coordinate distances between the starting point and preset grid vertices. This lays a reliable data foundation for subsequent grid-based historical trip matching and effectively improves the initial accuracy of destination prediction.

[0038] Furthermore, a trip database is constructed based on the historical trip trajectories of all private cars within the target area. Each historical trip trajectory in the database contains multi-dimensional trip information, such as departure time, date type, and trip trajectory sequence. The departure time is used to analyze periodic patterns; the date type is used to capture differentiated travel patterns of users on different dates, such as explicitly indicating whether the trip occurred on a weekday, weekend, or public holiday; and the trip trajectory sequence records a series of geographical coordinates, such as latitude and longitude, that the vehicle passed through during the trip in chronological order, in an ordered list format.

[0039] Before storing the original historical travel trajectories of private cars into the travel database, preprocessing is required. First, noise points in the trajectory sequence are removed. Original GPS and other positioning data often contain noise points that significantly deviate from the actual road due to signal obstruction, multipath effects, and other reasons. These points are identified and removed through methods such as speed, acceleration, or direction-based abrupt change detection, or road network-based matching filtering, ensuring that the trajectory sequence accurately reflects the vehicle's actual driving path. Second, the trajectory sequence is compressed. Specifically, the trajectory sequence after noise point removal is traversed, and the distance between two adjacent trajectory points (denoted as point A and point B) is calculated. If this distance is less than a preset distance threshold, point B is considered redundant and removed, while point A is retained and compared with the next point C. Because the compression threshold is influenced by both the grid size and regional characteristics—for example, a larger grid may allow for a longer tolerable compression distance, or the reasonable compression distance differs between highways and densely populated urban roads—the distance threshold in this embodiment is not a fixed value. It needs to be determined based on actual vehicle calibration. This threshold is determined by collecting data from actual vehicles and analyzing the balance between accuracy and efficiency, ensuring that the compression operation minimizes the amount of data while not losing feature points representing key steering and path characteristics. After preprocessing, the trajectory sequences of the historical travel trajectories of each private car are converted into the corresponding location codes mentioned above. Each trajectory point corresponds to a location code, which is the grid number. Thus, all continuous and complex trajectory points are standardized into discrete, unified grid number sequences, simplifying the complexity of subsequent trajectory matching, pattern recognition, and destination prediction calculations.

[0040] In one embodiment, since the starting point of a trip may not fall precisely at the center of a grid in real-world applications, and the starting point records of all private cars' historical trips within the target area also exhibit a certain spatial distribution, strictly matching only trip records whose starting points fall within the first grid might lead to the omission of a large number of spatially adjacent historical trips that can essentially be considered as originating from the same area. Therefore, this embodiment introduces a spatial redundancy query strategy, which involves filtering the historical trip set based on the first grid and its corresponding adjacent grids from the trip database. Adjacent grids refer to the eight grids spatially adjacent to the first grid. Specifically, these adjacent grids, together with the first grid, form a 3x3 grid cluster, and the historical trip set is used for querying and filtering. The historical trip set refers to all historical travel route data whose starting points are recorded within these grids.

[0041] Furthermore, the departure time of the vehicle's current trip is obtained and matched with the departure time of each trip trajectory in the historical trip set. Matching refers to time window matching, for example, querying historical trip trajectories whose departure time differs from the current trip's by one hour. This further filters the historical trip set, which has been initially spatially selected, to identify first-line candidate trips that are highly similar to the vehicle's current trip query in terms of departure time patterns. This effectively balances query precision and recall, ensuring that the retrieved candidate trips are both spatially close to the current departure location.

[0042] For step S103, a correlation analysis is performed based on the first candidate route and the real-time location of the vehicle to obtain the target candidate route with respect to the vehicle.

[0043] In one embodiment, the high-precision latitude and longitude coordinates of the vehicle's starting position, collected in real time by a positioning device such as an onboard GPS at the start of the current trip, are used as the real-time starting coordinates of the current trip. Then, for each historical trip trajectory in the first candidate trip, the actual distance between its recorded starting coordinates and the real-time starting coordinates of the current trip is calculated. Each calculated actual distance is compared with a preset distance threshold. The setting of the distance threshold needs to comprehensively consider the density of the urban road network, the typical error range of the positioning device, and the accuracy requirements of the starting point in the actual application scenario. For example, in a densely populated urban road area, this threshold might be set to 500 meters, thus considering trips whose starting points are located within the same community or large block as related. All historical trips whose spatial distance between their starting points and the starting point of the current trip is less than this preset threshold are filtered out to construct the second candidate trip. This effectively filters out historical trips that, although matching in macroscopic dimensions such as time, have actual starting points that are far apart and may represent completely different travel routes, thereby significantly improving the accuracy and reliability of the overall prediction results.

[0044] In one embodiment, the real-time location of the vehicle under test is continuously acquired via GPS, which may include the vehicle's latitude and longitude coordinates, timestamp, and possible instantaneous speed and heading at a specific moment. After obtaining the real-time location, target candidate routes passing through a predetermined range around the real-time location are selected from a second candidate route pool. Specifically, a second grid corresponding to the real-time location is obtained. Based on a grid pre-established in the target area, the real-time location of the vehicle under test is mapped to the corresponding second grid. To avoid omissions due to grid boundary division or minor positioning drift, adjacent grids bordering the second grid are obtained from the target area.

[0045] Next, a third candidate route is queried from the second candidate routes that passes through the second grid and its corresponding adjacent grids. Specifically, the grid number contained in each historical route trajectory in the second candidate routes is queried. If the historical route trajectory contains the number of the second grid or any of its adjacent grids, it is considered that the historical route trajectory has passed through the area of ​​the second grid or any of its adjacent grids, and it is selected to construct the third candidate route. This process excludes historical routes that are not related to the vehicle's current position in terms of spatial path, thus improving the accuracy of the prediction results.

[0046] Furthermore, a circular electronic fence area is determined with the vehicle's current real-time location as the center and a preset distance threshold set when filtering the second candidate trip as the radius. For each historical trip trajectory in the third candidate trip, it is determined whether any part of the trajectory line formed by connecting consecutive trajectory points crosses this circular area. Trips whose trajectory lines intersect with this circular electronic fence area are identified as target candidate trips. Specifically, the following steps are included: Search for the second candidate itinerary associated with the departure location from the first candidate itinerary; Obtain the real-time location of the vehicle; Filter target candidate routes that pass through a set range around the real-time location from the second candidate routes.

[0047] Among them, querying the second candidate itinerary associated with the departure location from the first candidate itinerary includes: Get the origin of the first trip in the first candidate trip; Calculate the actual distance between the starting point and the departure position of the first journey; The first candidate trip whose actual distance is less than the preset distance threshold is determined as the second candidate trip.

[0048] Among them, the selection of target candidate trips from the second candidate trips that pass through a set range around the real-time location includes: Obtain the second grid corresponding to the real-time location; Obtain the adjacent grids corresponding to the second grid in the target area; From the second candidate routes, query the third candidate routes that pass through the second grid and the corresponding adjacent grids of the second grid; The set range is determined based on the real-time location as the center and a preset distance threshold as the radius; The third candidate trip that falls within the set range is identified as the target candidate trip.

[0049] In this embodiment, when screening the second candidate routes, the actual distance between the historical route starting point and the current departure position is calculated and compared with a threshold, achieving preliminary association based on precise spatial location and effectively filtering historical data with excessively large starting point deviations. Secondly, in the real-time location-based screening, a grid system is first used for rapid initial screening, significantly narrowing the calculation range. Then, a circular geometric region is used for precise spatial relationship judgment, ensuring matching accuracy. This embodiment utilizes a coarse-to-fine screening mechanism to effectively balance computational load and result accuracy, enabling the solution to maintain efficient and reliable operation in complex real-time road network environments. Ultimately, it provides a high-quality data foundation for downstream tasks such as vehicle navigation and route prediction.

[0050] For step S104, the median center is calculated using the destinations of the target candidate trips to obtain the predicted destination of the vehicle.

[0051] In one embodiment, the Weiszfeld algorithm is used to calculate the predicted final destination location. First, for several historical journey trajectories among the candidate journeys, the latitude and longitude coordinates of the journey endpoints of these historical trajectories are arithmetically averaged to obtain the arithmetic mean of the ordinate and the arithmetic mean of ... Calculate the arithmetic mean center of the trip endpoints, and calculate the weighting factor for each trip endpoint based on the arithmetic mean center; The predicted destination of the vehicle is obtained by weighting the trip endpoints based on weighting factors.

[0052] Specifically, the latitude and longitude coordinates of M destinations are extracted from the target candidate itineraries, and a two-dimensional point set is constructed based on these latitude and longitude coordinates. Each of these points The coordinates are ,in Represents longitude coordinates. Represents latitude coordinates.

[0053] For several historical travel trajectories in the target candidate itinerary, the latitude and longitude coordinates of the destination of these historical travel trajectories are arithmetically averaged to obtain the arithmetic mean of the ordinate and the arithmetic mean of ...

[0054]

[0055] in, These are the longitude coordinates of the initial center point estimate. These are the latitude coordinates of the initial center point estimate.

[0056] Next, set the number of iterations to... initial value =0, for the first In the next iteration, its current center point estimate for:

[0057] in, It is the first The longitude coordinates of the current center point estimate in the next iteration. It is the first The latitude coordinates of the current center point estimate in the nth iteration. And the ()th iteration is calculated using the following formula. ( ) times the center point estimate :

[0058] in, Is the ( The longitude coordinates of the current center point estimate in +1) iterations. Is the ( The latitude coordinates of the current center point estimate after +1) iterations.

[0059] For the In each iteration, the endpoint of each historical travel trajectory is determined. Calculate its value compared with the current center point estimate. Euclidean distance The reciprocal of the weighting factor :

[0060]

[0061] Regarding the estimated distance from the current center point The farther the destination Its weighting factor The smaller.

[0062] Based on the calculated weighting factors, all these journey endpoints are weighted and calculated to obtain the ( )th journey endpoint. ( ) times the center point estimate Longitude coordinates and latitude coordinates :

[0063]

[0064] Next, the change in Euclidean distance between the two iteration points is calculated. That is, the first The (nth iteration point and the ()th +1) iteration point:

[0065] like If the value is less than the preset convergence threshold, the Weiszfeld algorithm is considered to have converged, and the estimated center point value at this point is... This is the final predicted destination location. If not, continue with the iterative calculation.

[0066] Furthermore, during the iteration process, if the current center point estimate is infinitely close to or coincides with the endpoint of a historical travel trajectory, it will cause the Euclidean distance between them to increase. If the weights approach or equal to zero, they will tend to infinity, causing computational overflow. Therefore, to ensure the robustness of the algorithm, a very small smoothing factor can be introduced when calculating the weights. This is to prevent division by zero errors. Therefore, the formula for calculating the weighting factor can be optimized as follows:

[0067] This invention achieves more accurate and robust destination prediction by assigning weight factors to different trip endpoints and performing weighted calculations, ensuring that the final predicted destinations tend to cluster in the high-frequency core area of ​​historical trip endpoints, thereby effectively improving the accuracy and reliability of the prediction results.

[0068] In one embodiment of the present invention, after calculating the median center using the endpoints of the target candidate routes to obtain the predicted destination of the vehicle, a continuous monitoring loop is entered. Specifically, it continuously detects whether the real-time position of the vehicle has been updated. If so, it detects whether the second grid corresponding to the updated real-time position has changed. The second grid is the grid cell to which the real-time position of the vehicle is mapped according to a pre-established uniform grid coordinate system.

[0069] If the second grid changes, it means that the vehicle has moved from the grid where it was in the previous decision cycle to a new grid cell. Therefore, the destination prediction step needs to be re-executed based on the changed second grid. Based on the vehicle's latest grid position, adjacent grid queries and candidate route screening will be performed again, and finally, the new predicted destination will be calculated using the Weiszfeld algorithm. This ensures that the prediction model can keep up with the actual driving trajectory of the vehicle and achieve dynamic updating and calibration of the prediction results.

[0070] If the second grid remains unchanged, it indicates that although the vehicle has moved, its spatial location remains within the same basic geographic unit (i.e., the same grid). In this case, the current destination prediction result is maintained. Simultaneously, to avoid the vehicle lingering or parking near the destination for an extended period, this embodiment starts a timer, terminating the prediction after a preset duration of no change in the second grid. The preset duration is a parameter that needs to be calibrated based on the actual scenario, for example, it can be set to 5 minutes or 10 minutes. On the one hand, this avoids unnecessary restarts of the prediction process when the vehicle stops briefly, such as while waiting at a red light; on the other hand, when the vehicle remains within the same grid for an extended period, it can intelligently determine that the trip status has stabilized or ended, thereby proactively terminating the prediction and releasing system computing resources. Specifically, the steps include the following: Detect whether the vehicle's real-time location has been updated; If the real-time location is updated, check whether the second grid corresponding to the updated real-time location has changed; If the second grid changes, the destination prediction steps are re-executed based on the changed second grid; or, if the second grid does not change, the current destination prediction result is maintained until the prediction is terminated after a preset duration in which the second grid remains unchanged.

[0071] This invention continuously monitors vehicle position and grid changes to capture deviations in the vehicle's trajectory. When the trajectory changes significantly, the destination is re-predicted, ensuring the prediction results are always synchronized with the latest driving status. It not only fully utilizes real-time location information throughout the journey to continuously correct predictions but also avoids unnecessary computational overhead through a delayed termination mechanism when the grid remains unchanged. This optimizes resource utilization while ensuring prediction accuracy and real-time performance.

[0072] This embodiment also provides a destination location prediction device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0073] This embodiment provides a destination location prediction device, such as... Figure 3 As shown, the device includes: The data acquisition module 301 is used to acquire the vehicle's departure location and departure time.

[0074] The trip query module 302 is used to obtain the first grid corresponding to the departure location in the target area, and query the first candidate trip that matches the departure time based on the historical trip set involved in the first grid and the adjacent grids corresponding to the first grid. The target area is the geographical area covered by the vehicle's travel, and the target area includes multiple evenly divided grids.

[0075] The trip analysis module 303 is used to perform correlation analysis between the first candidate trip and the real-time location of the vehicle to obtain the target candidate trip related to the vehicle.

[0076] The location prediction module 304 is used to calculate the median center using the destination of the target candidate trip to obtain the predicted destination of the vehicle.

[0077] The itinerary query module 302 includes: The mesh vertex acquisition module is used to obtain the vertex position of any mesh vertex in the target area.

[0078] The horizontal and vertical distance calculation module is used to calculate the horizontal and vertical distances of the starting position relative to the vertex position.

[0079] The first grid determination module is used to determine the first grid into which the starting position falls by utilizing the horizontal and vertical distances.

[0080] Trip analysis module 303 includes: The itinerary query module is used to query the second candidate itinerary associated with the departure location from the first candidate itinerary.

[0081] The real-time location acquisition module is used to acquire the real-time location of the vehicle.

[0082] The trip filtering module is used to filter target candidate trips that pass through a set range around the real-time location from the second candidate trips.

[0083] The itinerary query module includes: The trip endpoint acquisition unit is used to acquire the first trip start point in the first candidate trip.

[0084] The actual distance calculation unit is used to calculate the actual distance between the starting point of the first journey and the starting position.

[0085] The second candidate route determination unit is used to determine the first candidate route whose actual distance is less than a preset distance threshold as the second candidate route.

[0086] The itinerary filtering module includes: The second grid acquisition unit is used to acquire the second grid corresponding to the real-time location.

[0087] The adjacent grid acquisition unit is used to acquire the adjacent grids corresponding to the second grid from the target area.

[0088] The candidate route query unit is used to query the third candidate route from the second candidate route that passes through the second grid and the corresponding adjacent grid of the second grid.

[0089] The set range determination unit is used to determine the set range based on the real-time location as the center and the preset distance threshold as the radius.

[0090] The target candidate route determination unit is used to determine the third candidate route that falls within the set range as the target candidate route.

[0091] Location prediction module 304 includes: The weighting calculation unit is used to calculate the arithmetic mean center of the journey endpoints and to calculate the weighting factor for each journey endpoint based on the arithmetic mean center.

[0092] The weighted prediction unit is used to perform weighted calculations on the trip destination based on weighting factors to obtain the predicted destination of the vehicle.

[0093] Following the location prediction module 304, the device also includes: The location update module is used to detect whether the vehicle's real-time location has been updated.

[0094] The grid detection module is used to detect whether the second grid corresponding to the updated real-time position has changed if the real-time position is updated.

[0095] The prediction decision module is used to re-execute the destination prediction step based on the changed second grid if the second grid changes; or, if the second grid does not change, maintain the current destination prediction result until the prediction is terminated after a preset duration in which the second grid does not change.

[0096] In this embodiment, the destination location prediction device is presented in the form of a functional unit. Here, a unit refers to an ASIC circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above-mentioned functions.

[0097] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0098] This invention also provides a computer device having the above-described features. Figure 3 The device shown is a prediction device for the destination location.

[0099] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 4 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 4 Take a processor 10 as an example.

[0100] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0101] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0102] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0103] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0104] Input device 30 can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the computer device, such as a touchscreen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc. Output device 40 may include display devices, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors). The aforementioned display devices include, but are not limited to, liquid crystal displays, light-emitting diodes, displays, and plasma displays. In some alternative embodiments, the display device may be a touchscreen.

[0105] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0106] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0107] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for predicting the location of a destination, characterized in that, The method includes: Obtain the vehicle's departure location and departure time; Obtain the first grid corresponding to the departure location in the target area, and query the first candidate trip that matches the departure time based on the historical trip set involved in the first grid and the adjacent grids corresponding to the first grid. The target area is the geographical area covered by the vehicle's travel, and the target area includes multiple uniformly divided grids. Based on the correlation analysis between the first candidate route and the real-time location of the vehicle, a target candidate route with respect to the vehicle is obtained; The predicted destination of the vehicle is obtained by calculating the median center using the endpoints of the target candidate routes.

2. The method according to claim 1, characterized in that, The step of obtaining the first grid corresponding to the starting position in the target area includes: Obtain the vertex position of any grid vertex in the target region; Calculate the lateral and longitudinal distances of the starting position relative to the vertex position; Using the lateral distance and the longitudinal distance, the first grid into which the starting position falls is determined.

3. The method according to claim 1, characterized in that, The step of performing real-time correlation analysis based on the first candidate route and the real-time location of the vehicle to obtain the target candidate route with respect to the vehicle includes: Query the first candidate itinerary for a second candidate itinerary associated with the departure location; Obtain the real-time location of the vehicle; Select target candidate routes from the second candidate routes that pass through a set range around the real-time location.

4. The method according to claim 3, characterized in that, The step of querying a second candidate itinerary associated with the departure location from the first candidate itinerary includes: Obtain the starting point of the first trip in the first candidate trip; Calculate the actual distance between the starting point of the first journey and the starting position; The first candidate route whose actual distance is less than a preset distance threshold is determined as the second candidate route.

5. The method according to claim 3, characterized in that, The step of filtering target candidate trips from the second candidate trips that pass through a set range around the real-time location includes: Obtain the second grid corresponding to the real-time location; Obtain the adjacent grids corresponding to the second grid from the target region; From the second candidate routes, query the third candidate routes that pass through the second grid and the corresponding adjacent grids of the second grid; The set range is determined based on the real-time location as the center and the preset distance threshold as the radius; The third candidate route that falls within the set range is determined as the target candidate route.

6. The method according to claim 1, characterized in that, The step of calculating the median center using the destinations in the target candidate trips to obtain the predicted destination of the vehicle includes: Calculate the arithmetic mean center of the trip endpoints, and calculate the weighting factor for each trip endpoint based on the arithmetic mean center; The predicted destination of the vehicle is obtained by weighting the trip endpoint according to the weighting factor.

7. The method according to claim 1, characterized in that, After calculating the median center using the destinations in the target candidate trips to obtain the predicted destination of the vehicle, the method further includes: Detect whether the real-time location of the vehicle has been updated; If the real-time position is updated, detect whether the second grid corresponding to the updated real-time position has changed; If the second grid changes, the destination prediction steps are re-executed based on the changed second grid; or, if the second grid does not change, the current destination prediction result is maintained until a preset duration after which the prediction is terminated.

8. A device for predicting the location of a destination, characterized in that, The device includes: The data acquisition module is used to obtain the vehicle's departure location and departure time; The trip query module is used to obtain the first grid corresponding to the departure location in the target area, and query the first candidate trip that matches the departure time based on the historical trip set involved in the first grid and the adjacent grids corresponding to the first grid. The target area is the geographical area covered by the vehicle's travel, and the target area includes multiple evenly divided grids. The trip analysis module is used to perform correlation analysis between the first candidate trip and the real-time location of the vehicle to obtain the target candidate trip related to the vehicle. The location prediction module is used to calculate the median center using the destinations in the target candidate routes to obtain the predicted destination of the vehicle.

9. A computer device, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the method of any one of claims 1 to 7.

Images