Parking area identification method and system based on road network analysis

Abstract

The invention provides a parking area identification method and system based on road network analysis, and relates to the technical field of parking area identification, and the method comprises the steps: constructing a multi-level road network structure model of a target area; mapping a real-time obtained result to nodes of a multi-level road network structure model; according to the regional position and the current incoming vector, the destination attraction attribute and the incoming access convenience attribute are propagated and superposed between the node and the road section along the multi-level road network structure model; carrying out local fusion and correction on the intention distribution field; calculating a parking difficulty index for the candidate parking area set, and calculating a walking distance index from the candidate parking area set to the target area at the same time; and executing parking adaptation screening, and outputting a personalized parking area recommendation list. According to the invention, the technical problem of low parking area identification accuracy based on road network analysis in the prior art can be solved, and the technical effect of improving the parking area identification accuracy is achieved.

Description

Technical Field

[0001] This application relates to the field of parking area identification technology, and in particular to a parking area identification method and system based on road network analysis. Background Technology

[0002] With the increasing complexity of urban road networks and the continuous rise in the number of motor vehicles, parking demand in urban travel is showing a significant upward trend. Users' preferences for parking location, convenience, and walking distance constantly change across different scenarios, requiring parking recommendation systems to consider multiple factors simultaneously, including road traffic characteristics, parking space availability, real-time traffic changes, and user preferences. Currently, existing parking recommendation methods mainly focus on static parking space information or rely on a single data source, such as sorting by the real-time availability of parking spaces or by distance to the destination, neglecting issues such as road congestion, route complexity, and local traffic bottlenecks that vehicles may encounter on their way to parking areas.

[0003] In summary, existing technologies suffer from several technical problems. These include a lack of multi-dimensional traffic data fusion, a lack of a comprehensive quantitative mechanism for parking difficulty, a lack of in-depth modeling of user dynamic preferences, and a lack of a recommendation list optimization mechanism based on backtracking behavior. As a result, the accuracy and practicality of parking recommendation results are insufficient, which further affects the user's parking experience, vehicle traffic efficiency, and the overall operational quality of the urban transportation system under high load conditions. Summary of the Invention

[0004] The purpose of this application is to provide a parking area identification method and system based on road network analysis, in order to solve the technical problems in the existing technology, which are due to the lack of multi-dimensional traffic data fusion, the lack of a comprehensive quantitative mechanism for parking difficulty, the lack of deep modeling of user dynamic preferences, and the lack of a recommendation list optimization mechanism based on backtracking behavior. These problems lead to insufficient accuracy and practicality of parking recommendation results, which further affect the user's parking experience, vehicle traffic efficiency, and the overall operational quality of the urban traffic system under high load conditions.

[0005] In view of the above problems, this application provides a parking area identification method and system based on road network analysis.

[0006] Firstly, this application provides a parking area identification method based on road network analysis, implemented through a parking area identification system based on road network analysis. The method includes: after acquiring a target area, constructing a multi-level road network structure model of the target area; acquiring the vehicle's current position, historical trajectory segments, and current direction vector in real time, mapping the real-time acquisition results to nodes of the multi-level road network structure model to form a context mapping result; based on the context mapping result, propagating along the multi-level road network structure model between nodes and road segments according to the target area's location and current direction vector, and superimposing destination attraction attributes and direction access convenience attributes to construct an intent distribution field; after locally fusing and correcting the intent distribution field, outputting an intent aggregation area; using the intent aggregation area to construct a candidate parking area set, calculating a parking difficulty index for the candidate parking area set based on road segment complexity, historical occupancy rate, and playback simulation results, and simultaneously calculating a walking distance index from the candidate parking area set to the target area; after reading the user's parking preference information, performing parking adaptation filtering based on the parking difficulty index and walking distance index, and outputting a personalized parking area recommendation list.

[0007] Preferably, the parking area identification method based on road network analysis further includes: the multi-level road network structure model is constructed by labeling the road network with multi-dimensional attributes such as road level, node turning complexity, road segment traffic directionality, road segment accessibility, and parking availability, which is used to characterize the destination accessibility and incoming access convenience of the road network.

[0008] Preferably, the parking area identification method based on road network analysis further includes: reading historical traffic data and real-time traffic data, extracting traffic flow fluctuation characteristics, road segment congestion degree characteristics, and historical vehicle entry and exit statistics, and establishing an extracted feature set; dynamically constructing propagation weights for destination attraction attributes and inbound access convenience attributes based on the extracted feature set, real-time driving status, and current inbound vector; applying the propagation weights to the node and road segment propagation process in the multi-level road network structure model, gradually propagating the destination attraction attributes and inbound access convenience attributes in the multi-level road network structure model, and dynamically adjusting the propagation range and propagation intensity based on the real-time traffic data; performing traffic flow and congestion feature identification during the propagation process, constructing a local optimized response based on the identification results, performing local response optimization of the propagation, and constructing an intent distribution field based on the local response optimization results.

[0009] Preferably, the parking area identification method based on road network analysis further includes: using the historical traffic data and real-time traffic data as dual input data, and utilizing a multi-dimensional spatiotemporal traffic prediction channel to estimate the traffic flow fluctuations, traffic congestion risks, and potential traffic bottlenecks of each road segment in the multi-level road network structure model in real time, and establishing a time-series estimation result set; using the time-series estimation result set to analyze the real-time traffic status and historical congestion patterns of road segments and nodes, identifying the time periods and road segments where possible traffic congestion bottlenecks or traffic events may occur, and establishing identification results; adjusting the path propagation weights of local paths according to the identification results, and performing local response optimization management.

[0010] Preferably, the parking area identification method based on road network analysis further includes: constructing a parking access path based on the current location and the set of parking areas; extracting features from the parking access path such as the number of lanes, traffic light density, number of intersections, degree of turning, real-time traffic data, and road segment length to construct an accessibility complexity index; obtaining historical parking occupancy rate data from the candidate parking area set, the historical parking occupancy rate data including parking occupancy data with time period and date identifiers; fitting occupancy trends based on the historical parking occupancy rate data, using the occupancy trend fitting results to predict pre-parking saturation, and establishing a predicted parking complexity index; and calculating a parking difficulty index based on the accessibility complexity index and the predicted parking complexity index.

[0011] Preferably, the parking area identification method based on road network analysis further includes: reading the user's vehicle data, the vehicle data including vehicle size and model parameters; obtaining parking space data for each parking space in the parking area set; performing spatial adaptability analysis based on the parking space data and vehicle data to establish a spatial complexity index; extracting parking space type, parking space layout, and parking space location from the parking space data; using the extraction results and vehicle size to perform location layout complexity analysis to establish an adaptation complexity index; and calculating a parking difficulty index based on the spatial complexity index, adaptation complexity index, accessibility complexity index, and predicted parking complexity index.

[0012] Preferably, the parking area identification method based on road network analysis further includes: calling the mapped parking space camera according to the parking space data, performing parking space detection, and extracting additional features of the parking space, the additional features of the parking space including obstacle features inside the parking space and empty space features outside the parking space; and calculating the parking difficulty index after performing adaptation complexity index correction using the additional features of the parking space.

[0013] Preferably, the parking area identification method based on road network analysis further includes: parsing the parking preferences, constructing a preference feature set, the preference feature set including free preference priority, parking space preference, security management preference, walking accessibility preference, and operation difficulty preference; after converting the preference feature set into quantitative balance indicators, performing parking adaptation screening based on the parking difficulty indicator and walking distance indicator, and outputting a personalized parking area recommendation list.

[0014] Preferably, the parking area identification method based on road network analysis further includes: reading the user's preset window's retrospective selection data; reconstructing the personalized parking area recommendation list based on the retrospective selection data to establish a candidate list; configuring recommendation notifications for the candidate list, and managing the recommendation display of the candidate list after the user selects the corresponding recommendation notification.

[0015] Secondly, this application also provides a parking area identification system based on road network analysis, used to execute the parking area identification method based on road network analysis as described in the first aspect, including: a multi-level road network structure model construction module, used to construct a multi-level road network structure model of the target area after acquiring the target area; a context mapping result formation module, used to acquire the vehicle's current position, historical trajectory segments, and current direction vector in real time, and map the real-time acquisition results to the nodes of the multi-level road network structure model to form a context mapping result; and an intent distribution field construction module, used to construct the intent distribution field along the multi-level road network structure model between nodes and road segments based on the context mapping result and according to the regional position and current direction vector of the target area. The system propagates and overlays destination attraction attributes and access convenience attributes to construct an intent distribution field. An intent aggregation area output module outputs an intent aggregation area after local fusion and correction of the intent distribution field. A walking distance index calculation module constructs a candidate parking area set using the intent aggregation area, calculates a parking difficulty index for the candidate parking area set based on road segment complexity, historical occupancy rate, and replay simulation results, and simultaneously calculates the walking distance index from the candidate parking area set to the target area. A parking area recommendation list output module reads the user's parking preference information, performs parking adaptation filtering based on the parking difficulty index and walking distance index, and outputs a personalized parking area recommendation list.

[0016] The technical solution provided in this application has at least the following technical effects or advantages: by achieving the technical goals of globally modeling the complexity of the parking environment, dynamically characterizing user preferences, and intelligently optimizing the parking area matching logic, it achieves the technical effects of significantly improving the accuracy of parking recommendations, enhancing the adaptability of parking in different scenarios, and effectively improving the overall traffic operation efficiency.

[0017] The above description is merely an overview of the technical solution of this application. To enable a clearer understanding of the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

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

[0019] Figure 1 This is a flowchart illustrating the parking area identification method based on road network analysis proposed in this application.

[0020] Figure 2 This is a schematic diagram of the parking area identification system based on road network analysis in this application.

[0021] Figure labeling: Module 1 for constructing multi-level road network structure model, Module 2 for generating context mapping results, Module 3 for constructing intent distribution field, Module 4 for outputting intent concentration area, Module 5 for calculating walking distance index, and Module 6 for outputting parking area recommendation list. Detailed Implementation

[0022] This application provides a parking area identification method and system based on road network analysis. It addresses the technical problems in existing technologies, such as the lack of multi-dimensional traffic data fusion, a comprehensive quantitative mechanism for parking difficulty, deep modeling of dynamic user preferences, and a recommendation list optimization mechanism based on backtracking behavior. These shortcomings lead to insufficient accuracy and practicality of parking recommendations, further impacting user parking experience, vehicle traffic efficiency, and the overall operational quality of urban transportation systems under high load conditions. The application achieves the technical goals of globally modeling the complexity of the parking environment, dynamically characterizing user preferences, and intelligently optimizing parking area matching logic. This results in significantly improved parking recommendation accuracy, enhanced parking adaptability in different scenarios, and effective improvement of overall traffic efficiency.

[0023] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0024] Example 1, please refer to the appendix. Figure 1 This application provides a parking area identification method based on road network analysis, which is applied to a parking area identification system based on road network analysis, and specifically includes the following steps: S1: After obtaining the target area, construct a multi-level road network structure model of the target area.

[0025] Specifically, the scope of the urban area or street that the user is interested in is determined, and then a multi-level road network structure model that can reflect different levels of traffic structure is established within the scope of the urban area or street that the user is interested in. The multi-level refers to the hierarchical division of different road levels such as urban arterial roads, secondary arterial roads, and branch roads. At the same time, the branching relationship, intersection structure and hierarchical dependency relationship between each road will also be included in the model, so that the subsequent analysis can be closer to the real traffic environment.

[0026] S2: Real-time acquisition of the vehicle's current location, historical trajectory segments, and current direction vector, mapping the real-time acquisition results to the nodes of the multi-level road network structure model to form a context mapping result.

[0027] Specifically, the real-time acquisition of the vehicle's current position, historical trajectory segments, and current direction vector means continuously reading multiple data related to the vehicle's movement during the vehicle's operation. The current position refers to the vehicle's precise geographic coordinates at a certain moment, the historical trajectory segments refer to the vehicle's movement path segments over a period of time, used to reflect the vehicle's previous driving trend, and the current direction vector refers to the vector expression of the vehicle's current direction of movement, which can describe the direction from which the vehicle is approaching the target area.

[0028] Next, the real-time acquired results are mapped to nodes in the multi-level road network structure model. This means mapping real-time data to specific nodes in the model, projecting the vehicle's actual location onto the nearest road point or intersection. This transforms the vehicle from a point on the map into a valid location within the multi-level road network structure model. Simultaneously, the vehicle's travel direction and historical trajectory are used to determine the most likely road segment it is on, thus establishing a connection between the vehicle's state and the multi-level road network structure model. Finally, the resulting context mapping means that after mapping, we not only know the node the vehicle is on, but also understand its direction of origin, the roads it has traveled along, and the area it might be heading to.

[0029] S3: Based on the context mapping results, according to the regional location of the target area and the current direction vector, propagate along the multi-level road network structure model between nodes and road segments and superimpose the destination attraction attribute and the direction access convenience attribute to construct the intention distribution field.

[0030] Specifically, based on the context mapping results, according to the target area's location and current arrival vector (i.e., the vehicle's current location, historical trajectory fragments, and the corresponding node of the current arrival vector in the road network), the intention distribution field is constructed by propagating along a multi-level road network structure model between nodes and road segments and superimposing destination attraction attributes and arrival access convenience attributes. Here, "regional location" refers to the spatial location of the target area in the road network, such as its distance from main roads or whether it is in a traffic-intensive area; "current arrival vector" refers to the direction in which the vehicle is approaching the area, used to determine which roads the vehicle is more likely to enter. Subsequently, propagation occurs within the multi-level road network structure model, which refers to a hierarchical road network structure built according to road level, directionality, and accessibility, allowing different roads to have different influence ranges during propagation. During propagation, two types of attributes are gradually superimposed between nodes and road segments: the destination attraction attribute reflects the likelihood of a certain area being a vehicle's destination, for example, a commercial area may be more attractive than a residential area; the arrival access convenience attribute reflects the ease with which a vehicle enters a road segment from the current direction, for example, a left-turn route may be less convenient than a straight route. The superposition creates a structure that spatially represents the probability distribution of a vehicle's potential intentions, known as the intention distribution field, which describes the areas where a vehicle is more likely to move.

[0031] S4: After local fusion and correction of the intent distribution field, output the intent concentration region.

[0032] Specifically, after local fusion and correction of the intent distribution field, an intent-intensive region is output. Local fusion refers to integrating the intent intensity of adjacent areas, adjacent road segments, or interconnected nodes in the distribution field, smoothing out scattered high or low values ​​that might have been caused by noise or instantaneous traffic fluctuations. For example, if two adjacent road segments on a main road have intent values ​​of 60% and 65% respectively, the fusion may form a more stable 62% region to reduce unnecessary jumps. Correction adjusts the propagated intent values ​​based on real-time traffic, path accessibility, or historical patterns. For example, if an area has a high intent value but real-time traffic is congested, the correction mechanism will reduce its intent intensity to make the result more consistent with the actual road conditions. The final output intent-intensive region refers to the area where intents show a spatial concentration trend after fusion and correction, representing the most likely destination direction or target segment for vehicles, used for subsequent parking area selection.

[0033] S5: Construct a set of candidate parking areas using the intent-intensive area, calculate the parking difficulty index of the candidate parking area set based on road segment complexity, historical occupancy rate, and playback simulation results, and simultaneously calculate the walking distance index from the candidate parking area set to the target area.

[0034] Specifically, an intent-focused region is used to construct a set of candidate parking areas. The intent-focused region is used to infer the possible destination range of a vehicle; the set of candidate parking areas refers to multiple potentially available parking locations selected within the intent-focused region, such as public parking lots, temporary parking areas, or roadside parking strips. By associating vehicle intent with parking resources, the selection of candidate parking areas becomes more targeted.

[0035] Next, parking difficulty indices are calculated for the candidate parking area set based on road segment complexity, historical occupancy rates, and replay simulation results. Road segment complexity refers to the difficulty of the path from the vehicle's current location to the candidate parking area, including road width, number of traffic lights, number of intersections, curvature changes, and congestion conditions. Historical occupancy rate refers to the proportion of parking spaces occupied in the parking area on different dates and times in the past; for example, a parking lot may have an occupancy rate of over 90% at 7 PM every day. Replay simulation results refer to parking scenario prediction simulations generated based on historical traffic data and vehicle behavior data, such as simulating parking congestion after vehicles arrive in the area at the same time over the past 10 days. The comprehensive calculation of parking difficulty indices ensures that the final result reflects structural, behavioral, and predictive levels of difficulty.

[0036] At the same time, the walking distance metric from the candidate parking area set to the target area is calculated. The target area refers to the user's actual destination, such as an office building entrance, a shopping mall main entrance, or a residential unit. The walking distance metric represents the distance from each candidate parking area to the target area, which may include factors affecting the walking experience, such as stairs, pedestrian crossings, or pedestrian ramps. It can measure parking convenience, so that the final recommendation results take into account both parking difficulty and user walking comfort.

[0037] S6: After reading the user's parking preference information, perform parking adaptation filtering based on the parking difficulty index and walking distance index, and output a personalized parking area recommendation list.

[0038] Specifically, after reading the user's parking preference information, the system performs parking adaptation filtering based on parking difficulty and walking distance indicators to generate recommendations that better match the user's habits. Parking preference information refers to a user's specific tendencies in parking behavior, such as whether they prefer wide parking spaces, prioritize walking distance, desire safe parking lots, or prefer parking with reduced reversing difficulty. Parking difficulty indicators are calculated by the system based on road complexity, parking space structure, and traffic conditions. For example, a route with five traffic lights and three sharp turns is more difficult than one with only one traffic light. Walking distance indicators refer to the walking distance from the parking area to the target area. For example, one parking lot is 100 meters from the mall entrance, while another is 350 meters. By matching these indicators with user preferences—for example, if the user prioritizes walking distance highly, parking areas with shorter walking distances are recommended; if the user prefers ease of operation, the weight of areas with high parking difficulty is reduced. Finally, a personalized parking area recommendation list is output based on the filtering results, containing multiple parking options that meet user preferences and environmental constraints.

[0039] Furthermore, this application also includes: the multi-level road network structure model is constructed by labeling the road network with multi-dimensional attributes such as road level, node turning complexity, road segment traffic directionality, road segment accessibility, and parking availability, which is used to characterize the destination accessibility and incoming access convenience of the road network.

[0040] Specifically, the multi-level road network structure model constructs multi-dimensional attributes of the road network through road grade, node turning complexity, road segment traffic directionality, road segment accessibility, and parking availability. This means that during the modeling process, multiple dimensions of attribute information are added to each road segment and each node. For example, road grade distinguishes the traffic capacity of different roads, node turning complexity reflects the operational difficulty of vehicles turning at intersections, road segment traffic directionality describes the characteristics of one-way or two-way traffic, road segment accessibility reflects the strength of a road segment's ability to reach its destination, and parking availability identifies the area surrounding a road segment with available parking spaces. Then, the multi-level road network structure model is used to characterize the road network's destination accessibility and ease of access from different directions, i.e., the ease with which vehicles enter the target area from different directions, and the overall ease or difficulty of vehicles reaching the target location from a certain road. This helps determine whether a particular road segment is more suitable as a guiding path for a potential parking spot for the driver.

[0041] Furthermore, this application also includes: reading historical traffic data and real-time traffic data, extracting traffic flow fluctuation characteristics, road segment congestion degree characteristics, and historical vehicle entry and exit statistics, and establishing an extracted feature set; dynamically constructing propagation weights for destination attraction attributes and inbound access convenience attributes based on the extracted feature set, real-time driving status, and current inbound vector; applying the propagation weights to the node and road segment propagation process in the multi-level road network structure model, gradually propagating the destination attraction attributes and inbound access convenience attributes in the multi-level road network structure model, and dynamically adjusting the propagation range and propagation intensity based on the real-time traffic data; performing traffic flow and congestion feature identification during the propagation process, constructing a local optimized response based on the identification results, performing local response optimization of the propagation, and constructing an intent distribution field based on the local response optimization results.

[0042] Specifically, historical and real-time traffic data are read to extract traffic flow fluctuation characteristics, road segment congestion characteristics, and historical vehicle entry and exit statistics. An extracted feature set is then established, representing the loading of data from both sources. Historical traffic data describes road traffic conditions at different times in the past, while real-time traffic data reflects the current actual traffic flow and congestion status on the road. Traffic flow fluctuation characteristics refer to the dynamic pattern of vehicle numbers changing over time, such as the difference in traffic flow during morning and evening rush hours; road segment congestion characteristics refer to the severity of congestion at different times; and historical vehicle entry and exit statistics describe the distribution of vehicle entry and exit numbers in a specific area or road segment during a typical time period.

[0043] Next, based on the extracted feature set, real-time driving status, and current approach vector, the propagation weights of the destination attraction attribute and the approach access convenience attribute are dynamically constructed. This explains how the feature set, combined with real-time status information such as the vehicle's current speed, acceleration, and direction, as well as the approach direction indicated by the approach vector, is used to generate the propagation weights for these two key attributes. The destination attraction attribute refers to the degree to which certain areas are more likely to become the vehicle's destination, while the approach access convenience attribute refers to the ease with which the vehicle can access certain roads or areas from its current direction. The dynamic construction of the propagation weights means that they are adjusted in real time according to changes in traffic conditions; for example, when congestion on a road worsens, the propagation weight of its attraction attribute will decrease.

[0044] Subsequently, propagation weights are applied to the node and road segment propagation process in the multi-level road network structure model. Destination attraction and inbound access convenience attributes are gradually propagated within the multi-level road network structure model, and the propagation range and intensity are dynamically adjusted based on real-time traffic data. This means performing a diffusion-like propagation process within the road network structure, spreading the propagation weights from a road or node to the surrounding area to reflect the accessibility connections between different roads. The propagation range refers to the road levels and distances that the propagation can reach, while the propagation intensity represents the degree to which an attribute attenuates or strengthens during propagation. Both are affected by real-time traffic conditions; for example, increased traffic volume may reduce the propagation range, while smooth traffic flow may increase the propagation intensity.

[0045] Finally, the propagation process identifies traffic flow and congestion characteristics. Based on the identification results, a locally optimized response is constructed, and the local response optimization is performed. Based on the optimized local response results, an intent distribution field is constructed, representing the continuous monitoring of traffic flow and congestion on each road during the propagation process, identifying potential bottlenecks, such as detecting sudden congestion on a main road. Based on the identification results, the propagation mechanism is locally optimized, for example, by reducing the propagation weight of congested areas or increasing the diffusion influence of unobstructed roads, to ensure that the overall propagation result better reflects real traffic trends. The final optimized propagation structure forms an intent distribution field, representing the probability distribution of possible vehicle intention directions and potential destination areas.

[0046] Furthermore, this application also includes: using the historical traffic data and real-time traffic data as dual input data, and utilizing a multi-dimensional spatiotemporal traffic prediction channel to estimate the traffic fluctuations, traffic congestion risks, and potential traffic bottlenecks of each road segment in the multi-level road network structure model in real time, and establishing a time-series estimation result set; using the time-series estimation result set to analyze the real-time traffic status and historical congestion patterns of road segments and nodes, identifying the time periods and road segments where possible traffic congestion bottlenecks or traffic events may occur, and establishing identification results; adjusting the path propagation weights of local paths according to the identification results, and performing local response optimization management.

[0047] Specifically, using historical and real-time traffic data as dual inputs, a multi-dimensional spatiotemporal traffic prediction channel is employed to estimate traffic flow fluctuations, congestion risks, and potential bottlenecks for each road segment in a multi-level road network structure model in real time. This results in a time-series estimation set, representing the use of both historical and real-time data as inputs to more accurately predict traffic changes in both time and space dimensions. The multi-dimensional spatiotemporal traffic prediction channel is a prediction module capable of simultaneously processing time-series and spatial correlation features to estimate vehicle flow changes on each road in the near future. Traffic flow fluctuations describe the increase or decrease in the number of vehicles over a period of time; congestion risk indicates whether a road segment is likely to become congested in the near future; and potential bottlenecks refer to key locations where traffic flow is concentrated and prone to congestion. All prediction results are integrated to form a time-series estimation set, used to express the potential trends in traffic conditions in the near future.

[0048] Subsequently, the real-time traffic conditions and historical congestion patterns of road segments and nodes are analyzed using the time-series estimation results set. This identifies the time periods and road segments where potential traffic bottlenecks or traffic events may occur, establishing identification results. This demonstrates that by analyzing the predicted data, combining the current real-time traffic conditions with past congestion patterns occurring in similar time periods, areas in the road network that may experience traffic problems can be identified. Traffic bottlenecks refer to critical road segments where traffic flow is forced to slow down or queue, while traffic events can include minor accidents, temporary construction, or traffic light malfunctions, which may lead to sudden changes in traffic conditions. The identification results mark which road segments and at what time periods these events are more likely to occur, thus providing a basis for subsequent optimization.

[0049] Next, based on the identification results, the path propagation weights of local paths are adjusted, and local response optimization management is performed. This means that based on the identified potential bottlenecks and local congestion risks, the weights in the propagation mechanism are locally adjusted, reducing the influence of areas in intent propagation or increasing the weights of alternative routes, so that the overall predicted path better reflects actual road accessibility. The role of local response optimization management is to enable the propagation model to automatically adapt to dynamically changing traffic conditions, making vehicle intent inference and route analysis more accurate.

[0050] Furthermore, this application also includes: constructing a parking access path based on the current location and the set of parking areas; extracting features from the parking access path such as the number of lanes, traffic light density, number of intersections, degree of turning, real-time traffic data, and road segment length to construct an accessibility complexity index; obtaining historical parking occupancy rate data from the candidate parking area set, the historical parking occupancy rate data including parking occupancy data with time period and date identifiers; fitting occupancy trends based on the historical parking occupancy rate data, using the occupancy trend fitting results to predict pre-parking saturation, and establishing a predicted parking complexity index; and calculating a parking difficulty index based on the accessibility complexity index and the predicted parking complexity index.

[0051] Specifically, parking routes are constructed based on the current location and the set of parking areas. Starting from the current location, possible routes to each parking area are found according to the road network structure and traffic rules. For example, if the current location is in the middle of a road segment and there are multiple parking lots nearby, multiple candidate routes from the current location to the nearby parking lots will be generated.

[0052] Next, features are extracted for each parking route, including the number of lanes, traffic light density, number of intersections, degree of curvature, real-time traffic data, and route length. The number of lanes refers to the number of lanes available for vehicles traveling in the same direction, traffic light density represents the number of traffic lights per unit distance, the number of intersections indicates the number of junctions along the route, degree of curvature describes the frequency and angular complexity of turns, real-time traffic data reflects current traffic speed and congestion status, and route length is the physical length of each segment. By extracting these features, an accessibility complexity index can be constructed to quantify driving difficulty, making the accessibility of different routes comparable.

[0053] Next, historical parking occupancy data is obtained from the candidate parking area set, including parking occupancy information with time and date identifiers, such as the proportion of parking spaces used in a parking lot on different dates and at different times. Historical parking occupancy rates can reflect the past busyness of parking areas, and time-labeled data can help analyze regular changes over different times.

[0054] Therefore, based on historical parking occupancy data, an occupancy trend fitting is performed. Trend fitting means using a mathematical model to approximate past occupancy changes as a curve, thereby predicting the possible future occupancy level. The results of the occupancy trend fitting can be used for pre-parking saturation prediction, that is, predicting whether a parking lot is likely to reach full capacity at a future point in time. Based on this, a predicted parking complexity index is established to quantify the future congestion level and parking difficulty of the parking lot.

[0055] Finally, the accessibility complexity index and the predicted parking complexity index are combined to calculate the parking difficulty index. The parking difficulty index numerically reflects the complete difficulty of getting from the current location to the parking area, including traffic complexity and the future availability of the parking area. This allows for a comprehensive evaluation of parking difficulty based on multiple factors.

[0056] Furthermore, this application also includes: reading the user's vehicle data, the vehicle data including vehicle size and model parameters; obtaining parking space data for each parking space in the parking area set, performing spatial adaptability analysis based on the parking space data and vehicle data, and establishing a spatial complexity index; extracting parking space type, parking space layout, and parking space location from the parking space data, performing location layout complexity analysis using the extraction results and vehicle size, and establishing an adaptation complexity index; and calculating a parking difficulty index based on the spatial complexity index, adaptation complexity index, accessibility complexity index, and predicted parking complexity index.

[0057] Specifically, this involves reading the user's vehicle data, which primarily includes vehicle dimensions and model parameters. Vehicle dimensions refer to the vehicle's length, width, height, and other spatial information, while model parameters refer to the vehicle's structural characteristics, such as whether it is a large SUV, a small sedan, or a new energy vehicle. These parameters affect the vehicle's maneuverability in confined spaces, its minimum turning radius, and its parking space requirements. Reading this data ensures that subsequent analysis is matched to the vehicle's inherent characteristics.

[0058] Next, parking space data for each parking location in the parking area set is acquired, and spatial adaptability analysis is performed based on the parking space data and vehicle data to establish a spatial complexity index. Parking space data refers to structural information such as the length, width, usable height, and gaps between parking spaces. Spatial adaptability analysis means assessing whether a vehicle can smoothly enter or park in a parking space without collision by comparing vehicle dimensions with parking space dimensions. The spatial complexity index is a quantitative indicator used to reflect the ease or difficulty of matching vehicle and parking space dimensions; for example, the narrower or shorter the parking space, the higher the spatial complexity index.

[0059] Subsequently, parking space types, layouts, and locations were extracted from the parking space data. The extracted results, combined with vehicle dimensions, were used to conduct a complexity analysis of the parking space layout to establish an adaptation complexity index. Parking space type refers to the form of the parking space, such as perpendicular, angled, or parallel parking spaces; parking space layout refers to the arrangement of parking spaces, such as whether they are arranged in multiple rows or whether the exit is convenient; parking space location refers to the specific location of the parking space within the entire parking lot, such as whether it is near a wall, pillar, or ramp entrance. The complexity analysis of the parking layout assesses the difficulty of parking under different layout and location conditions. For example, angled parking spaces are generally easier to park in than parallel parking spaces, while parking spaces near pillars are more difficult to park in due to more obstacles. The adaptation complexity index is then used to quantify the difficulty at the operational level.

[0060] Finally, the final parking difficulty index is calculated based on spatial complexity, adaptation complexity, accessibility complexity, and predicted parking complexity. The accessibility complexity index reflects the difficulty of a vehicle's path from its current location to the parking space, while the predicted parking complexity index represents the potential occupancy level of the parking space at a future point in time. By combining these indices, the parking difficulty index provides a more comprehensive assessment of parking difficulty, incorporating factors at the path level as well as those related to the parking space's structure and future availability.

[0061] Furthermore, this application also includes: calling the mapped parking space camera according to the parking space data, performing parking space detection, and extracting additional features of the parking space, the additional features of the parking space including obstacle features inside the parking space and empty space features outside the parking space; and calculating the parking difficulty index after performing adaptation complexity index correction using the additional features of the parking space.

[0062] Specifically, parking space data refers to the basic information of the parking space itself, such as its size, location, and layout. Mapped parking space cameras refer to monitoring devices bound to a specific parking space, capable of capturing real-time visual images of that space. The mapped parking space camera is invoked based on the parking space data, and parking space detection is performed to extract additional features. Performing parking space detection means identifying the current state of the parking space through image analysis technology, such as whether it is occupied or if there are obstructions. Additional parking space features refer to further information beyond the basic dimensions, including obstacle features within the parking space and features of the free space outside the parking space. Obstacle features within the parking space indicate whether there is trash, debris, pillars, or other structures that impede vehicle entry; features of the free space outside the parking space indicate the operable space around the parking space, such as whether there is sufficient space for reversing or turning, whether the space is adjacent to walls, or whether there are additional passageways or gaps.

[0063] Next, these additional parking space features are used to correct the adaptation complexity index, and finally, the parking difficulty index is calculated. The adaptation complexity index was originally calculated based on structural factors such as parking space size, type, and layout. The correction process means that the real-time physical environment detected by the camera is considered on top of the original assessment. For example, obstacles reduce adaptability, while increased free space increases it. The parking difficulty index is a final overall quantitative index derived from factors such as spatial complexity, adaptation complexity, accessibility complexity, and predicted parking complexity. The addition of additional parking space features makes the results closer to actual parking scenarios.

[0064] Furthermore, this application also includes: parsing the parking preferences, constructing a preference feature set, the preference feature set including free preference priority, parking space preference, security management preference, walking accessibility preference, and operation difficulty preference; after converting the preference feature set into quantitative balance indicators, performing parking adaptation screening based on the parking difficulty indicator and walking distance indicator, and outputting a personalized parking area recommendation list.

[0065] Specifically, user parking preferences are analyzed, and a preference feature set is constructed. Parking preferences refer to users' tendencies when choosing parking locations, such as whether they prefer free parking lots, whether they prefer more spacious parking spaces, whether they value security measures in parking areas, whether they want the shortest possible walking distance, or whether they prefer parking spaces with low operational difficulty. The preference feature set includes five categories of preferences: Free Parking Preference Priority refers to the user's emphasis on free parking; for example, some users may be willing to walk hundreds of meters more to save money. Parking Space Preference refers to whether users prefer spacious parking spaces; for example, users driving large SUVs may prefer parking spaces with a width of at least 2.8 meters. Security Management Preference refers to the user's emphasis on parking lot security measures; for example, whether they pay attention to the coverage of in-parking cameras or the intensity of nighttime lighting. Walkability Preference indicates the user's sensitivity to walking distance; for example, whether they prefer parking spaces less than 200 meters from the entrance. Operational Difficulty Preference represents whether users prefer parking spaces that are easy to enter or easy to reverse into, such as avoiding narrow corners or areas with dense pillars.

[0066] Next, the preference feature set is transformed into quantitative balance indicators, and adaptive filtering is performed on the parking difficulty indicator and walking distance indicator, ultimately outputting a personalized parking area recommendation list. Quantitative balance indicators refer to converting user preferences into calculable values ​​through mathematical weighting; for example, the weight of free preference is set to 0.4, and the weight of space preference is set to 0.2. The parking difficulty indicator reflects the difficulty of the parking route and parking space structure, while the walking distance indicator reflects the walking distance to the destination after parking. Indicators are filtered and sorted according to preference weights. For example, if the user has a high weight for walking accessibility, parking areas closer to the entrance are recommended first; if the user has a high preference for operational difficulty, parking spaces near pillars are avoided. The final output personalized parking area recommendation list contains multiple parking options that meet user preferences and actual environmental constraints.

[0067] Furthermore, this application also includes: reading the user's preset window's retrospective selection data; reconstructing the personalized parking area recommendation list based on the retrospective selection data to establish a candidate list; configuring recommendation notifications for the candidate list, and managing the recommendation display of the candidate list after the user selects the corresponding recommendation notification.

[0068] Specifically, the system reads the user's retrospective selection data within a preset window. The preset window refers to a time range set for recording user behavior, such as the last 7 days or the last 30 days. The retrospective selection data refers to the records of parking areas that the user actually selected, recommended items that they clicked, or abandoned selections during this period. It reflects the user's true behavioral preferences. For example, the user may often choose parking spaces near the entrance or areas that are paid but easier to park in. Therefore, the system can gain a deeper understanding of the user's long-term preference trends through retrospective records.

[0069] Next, the personalized parking area recommendation list is restructured based on the retrospective selection data, resulting in a candidate list that better reflects the user's historical preferences. List restructuring involves the system reordering or removing certain options from the original recommendation list. For example, if the system detects that a user has never selected a parking space more than 500 meters from the target area based on their past 10 selections, then such parking areas will be automatically downgraded during restructuring. Conversely, if a user has repeatedly selected wider parking spaces in the past, even if the parking fee is higher, wider parking spaces will be given higher priority in the candidate list. The candidate list thus becomes a comprehensive recommendation result that takes into account the real-time environment, personal preferences, and historical behavior.

[0070] Subsequently, the system configures recommendation notifications for the candidate list and manages the display of the candidate list after the user selects the notification. Recommendation notifications are push notifications to the user, such as pop-ups on mobile devices, in-vehicle screen reminders, or voice prompts, and may contain a summary of recommended parking areas. After the user selects a recommendation notification, the candidate list is presented in a more intuitive way, such as displaying the location of the parking area, walking distance, and estimated occupancy rate, thus helping the user make a quick parking decision. Recommendation display management refers to the system presenting this information with appropriate layout and priority based on the characteristics of the user's terminal interface or in-vehicle system.

[0071] In summary, the parking area identification method based on road network analysis provided in this application has the following technical effects: by achieving the technical goals of globally modeling the complexity of the parking environment, dynamically characterizing user preferences, and intelligently optimizing the parking area matching logic, it can significantly improve the accuracy of parking recommendations, enhance the adaptability of parking in different scenarios, and effectively improve the overall traffic operation efficiency.

[0072] Example 2: Based on the same inventive concept as the parking area identification method based on road network analysis in the previous examples, this application also provides a parking area identification system based on road network analysis. Please refer to the appendix. Figure 2 The system includes: a multi-level road network structure model construction module 1, used to construct a multi-level road network structure model of the target area after acquiring the target area; a context mapping result formation module 2, used to acquire the vehicle's current position, historical trajectory segments, and current direction vector in real time, and map the real-time acquisition results to the nodes of the multi-level road network structure model to form a context mapping result; and an intent distribution field construction module 3, used to construct an intent distribution field based on the context mapping result, according to the target area's regional location and current direction vector, by propagating along the multi-level road network structure model between nodes and road segments and superimposing destination attraction attributes and direction access convenience attributes. The system comprises: a field; an intent aggregation area output module 4, used to output an intent aggregation area after local fusion and correction of the intent distribution field; a walking distance index calculation module 5, used to construct a candidate parking area set using the intent aggregation area, calculate a parking difficulty index for the candidate parking area set based on road segment complexity, historical occupancy rate, and playback simulation results, and simultaneously calculate the walking distance index from the candidate parking area set to the target area; and a parking area recommendation list output module 6, used to perform parking adaptation filtering based on the parking difficulty index and walking distance index after reading the user's parking preference information, and output a personalized parking area recommendation list.

[0073] Furthermore, the parking area identification system based on road network analysis is also used to: construct the multi-dimensional attribute annotation of the road network through road level, node turning complexity, road segment traffic directionality, road segment accessibility, and parking availability, which is used to characterize the destination accessibility and incoming access convenience of the road network.

[0074] Furthermore, the parking area identification system based on road network analysis is also used for: reading historical traffic data and real-time traffic data, extracting traffic flow fluctuation characteristics, road segment congestion characteristics, and historical vehicle entry and exit statistics, and establishing an extracted feature set; dynamically constructing propagation weights for destination attraction attributes and inbound access convenience attributes based on the extracted feature set, real-time driving status, and current inbound vector; applying the propagation weights to the node and road segment propagation process in the multi-level road network structure model, gradually propagating the destination attraction attributes and inbound access convenience attributes in the multi-level road network structure model, and dynamically adjusting the propagation range and propagation intensity based on the real-time traffic data; performing traffic flow and congestion feature identification during the propagation process, constructing a local optimized response based on the identification results, performing local response optimization of the propagation, and constructing an intent distribution field based on the local response optimization results.

[0075] Furthermore, the parking area identification system based on road network analysis is also used to: take the historical traffic data and real-time traffic data as dual input data, utilize a multi-dimensional spatiotemporal traffic prediction channel to estimate the traffic flow fluctuations, traffic congestion risks, and potential traffic bottlenecks of each road segment in the multi-level road network structure model in real time, and establish a time-series estimation result set; use the time-series estimation result set to analyze the real-time traffic status and historical congestion patterns of road segments and nodes, identify the time periods and road segments where possible traffic congestion bottlenecks or traffic events may occur, and establish identification results; adjust the path propagation weights of local paths according to the identification results, and perform local response optimization management.

[0076] Furthermore, the parking area identification system based on road network analysis is also used to: construct a parking access path based on the current location and the set of parking areas; extract features from the parking access path, such as the number of lanes, traffic light density, number of intersections, degree of turning, real-time traffic data, and road segment length, to construct an accessibility complexity index; obtain historical parking occupancy rate data from the candidate parking area set, including parking occupancy data with time period and date identifiers; perform occupancy trend fitting based on the historical parking occupancy rate data, use the occupancy trend fitting results to predict pre-parking saturation, and establish a predicted parking complexity index; and calculate a parking difficulty index based on the accessibility complexity index and the predicted parking complexity index.

[0077] Furthermore, the parking area identification system based on road network analysis is also used for: reading the user's vehicle data, including vehicle size and model parameters; obtaining parking space data for each parking space in the parking area set; performing spatial adaptability analysis based on the parking space data and vehicle data to establish a spatial complexity index; extracting parking space type, parking space layout, and parking space location from the parking space data; using the extraction results and vehicle size to perform location layout complexity analysis to establish an adaptation complexity index; and calculating a parking difficulty index based on the spatial complexity index, adaptation complexity index, accessibility complexity index, and predicted parking complexity index.

[0078] Furthermore, the parking area recognition system based on road network analysis is also used to: call the mapped parking space camera according to the parking space data, perform parking space detection, extract additional features of the parking space, the additional features of the parking space include obstacle features inside the parking space and empty space features outside the parking space; and calculate the parking difficulty index after performing adaptation complexity index correction using the additional features of the parking space.

[0079] Furthermore, the parking area identification system based on road network analysis is also used to: parse the parking preferences, construct a preference feature set, the preference feature set including free preference priority, parking space preference, safety management preference, walking accessibility preference, and operation difficulty preference; after converting the preference feature set into quantitative balance indicators, perform parking adaptation screening based on the parking difficulty indicator and walking distance indicator, and output a personalized parking area recommendation list.

[0080] Furthermore, the parking area identification system based on road network analysis is also used to: read the user's preset window's retrospective selection data; reconstruct the personalized parking area recommendation list based on the retrospective selection data to establish a candidate list; configure the recommendation notification of the candidate list, and manage the recommendation display of the candidate list after the user selects the corresponding recommendation notification.

[0081] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The parking area identification method and specific examples based on road network analysis in the aforementioned embodiment 1 are also applicable to the parking area identification system based on road network analysis in this embodiment. Through the foregoing detailed description of the parking area identification method based on road network analysis, those skilled in the art can clearly understand the parking area identification system based on road network analysis in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0082] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0083] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A parking area identification method based on road network analysis, characterized in that, The method includes: After obtaining the target area, construct a multi-level road network structure model of the target area; The vehicle's current location, historical trajectory segments, and current direction vector are acquired in real time, and the real-time acquisition results are mapped to the nodes of the multi-level road network structure model to form a context mapping result. Based on the context mapping results, according to the regional location of the target area and the current direction vector, the intention distribution field is constructed by propagating along the multi-level road network structure model between nodes and road segments and superimposing the destination attraction attribute and the direction access convenience attribute. After local fusion and correction of the intent distribution field, the intent concentration region is output; A set of candidate parking areas is constructed using the intent-intensive area. The parking difficulty index is calculated for the set of candidate parking areas based on road segment complexity, historical occupancy rate, and playback simulation results. At the same time, the walking distance index from the set of candidate parking areas to the target area is also calculated. After reading the user's parking preference information, parking matching filtering is performed based on the parking difficulty index and walking distance index, and a personalized parking area recommendation list is output.

2. The parking area identification method based on road network analysis as described in claim 1, characterized in that, The multi-level road network structure model is constructed by labeling the road network with multi-dimensional attributes such as road level, node turning complexity, road segment traffic directionality, road segment accessibility, and parking availability, which are used to characterize the destination accessibility and the ease of access from the source of the road network.

3. The parking area identification method based on road network analysis as described in claim 1, characterized in that, Constructing an intent distribution field includes: Read historical and real-time traffic data, extract traffic flow fluctuation characteristics, road segment congestion characteristics, and historical vehicle entry and exit statistics, and establish an extracted feature set; Based on the extracted feature set, real-time driving status, and current direction vector, the propagation weights of the destination attraction attribute and the direction access convenience attribute are dynamically constructed. The propagation weights are applied to the nodes and road segments in the multi-level road network structure model during the propagation process. Destination attraction attributes and access convenience attributes are gradually propagated in the multi-level road network structure model, and the propagation range and intensity are dynamically adjusted according to the real-time traffic data. The process involves identifying traffic and congestion characteristics during the propagation process, constructing a local optimized response based on the identification results, optimizing the local response during propagation, and constructing an intent distribution field based on the local response optimization results.

4. The parking area identification method based on road network analysis as described in claim 3, characterized in that, The process involves identifying traffic and congestion characteristics during the propagation process, constructing a local optimal response based on the identification results, and performing local response optimization during propagation, including: Using the historical traffic data and real-time traffic data as dual input data, and utilizing the multi-dimensional spatiotemporal traffic prediction channel, the traffic fluctuations, traffic congestion risks, and potential traffic bottlenecks of each road segment in the multi-level road network structure model are estimated in real time, and a time-series estimation result set is established. The time-series estimation result set is used to analyze the real-time traffic status and historical congestion patterns of road segments and nodes, identify the time periods and road segments where possible traffic congestion bottlenecks or traffic events may occur, and establish identification results. Based on the identification results, the path propagation weights of the local paths are adjusted, and local response optimization management is performed.

5. The parking area identification method based on road network analysis as described in claim 1, characterized in that, For the candidate parking area set, a parking difficulty index is calculated based on road segment complexity, historical occupancy rate, and replay simulation results, including: Construct a parking access path based on the current location and the set of parking areas; The accessibility complexity index is constructed by extracting features such as the number of lanes, traffic light density, number of intersections, degree of turning, real-time traffic data, and road segment length from the parking and traffic routes. Obtain historical parking occupancy rate data from the candidate parking area set, wherein the historical parking occupancy rate data includes parking occupancy data with time period and date identifiers; Based on the historical parking occupancy rate data, an occupancy trend is fitted, and the occupancy trend fitting results are used to predict pre-parking saturation and establish a predictive parking complexity index. The parking difficulty index is calculated based on the accessibility complexity index and the predicted parking complexity index.

6. The parking area identification method based on road network analysis as described in claim 5, characterized in that, The parking difficulty index is calculated based on the accessibility complexity index and the predicted parking complexity index, including: Read the user's vehicle data, which includes vehicle dimensions and model parameters; Obtain parking space data for each parking space in the parking area set, perform spatial adaptability analysis based on the parking space data and vehicle data, and establish a spatial complexity index. Extract parking space type, parking space layout and parking space location from parking space data, use the extraction results and vehicle size to perform complexity analysis of location layout, and establish an adaptation complexity index. The parking difficulty index is calculated based on the space complexity index, adaptation complexity index, accessibility complexity index, and predicted parking complexity index.

7. The parking area identification method based on road network analysis as described in claim 6, characterized in that, The parking difficulty index is calculated based on the spatial complexity index, adaptation complexity index, accessibility complexity index, and predicted parking complexity index, including: Based on the parking space data, the mapped parking space camera is invoked to perform parking space detection and extract additional features of the parking space, including features of obstacles inside the parking space and features of empty space outside the parking space. After performing adaptation complexity index correction using the aforementioned additional features of the parking space, the parking difficulty index is calculated.

8. The parking area identification method based on road network analysis as described in claim 1, characterized in that, After reading the user's parking preference information, parking matching filtering is performed based on the parking difficulty index and walking distance index, and a personalized parking area recommendation list is output, including: The parking preferences are analyzed to construct a preference feature set, which includes free parking preference priority, parking space preference, safety management preference, walking accessibility preference, and operation difficulty preference. After converting the preference feature set into a quantitative balance index, the parking difficulty index and walking distance index are used for parking adaptation screening, and a personalized parking area recommendation list is output.

9. The parking area identification method based on road network analysis as described in claim 1, characterized in that, The output includes a personalized list of recommended parking areas, and also includes: Read the user's preset window's back selection data; Based on the backtracking selection data, the personalized parking area recommendation list is reconstructed to establish a candidate list; Configure the recommendation notification for the candidate list, and manage the recommendation display of the candidate list after the user selects the corresponding recommendation notification.

10. A parking area identification system based on road network analysis, characterized in that, The steps for implementing the parking area identification method based on road network analysis according to any one of claims 1 to 9 include: The multi-level road network structure model building module is used to build a multi-level road network structure model of the target area after the target area is obtained. The context mapping result generation module is used to acquire the vehicle's current position, historical trajectory segments, and current direction vector in real time, and map the real-time acquisition results onto the nodes of the multi-level road network structure model to form the context mapping result. The intent distribution field construction module is used to construct an intent distribution field based on the context mapping result, according to the regional location of the target area and the current direction vector, by propagating along the multi-level road network structure model between nodes and road segments and superimposing the destination attraction attribute and the direction access convenience attribute; The intent convergence area output module is used to output the intent convergence area after local fusion and correction of the intent distribution field; The walking distance index calculation module is used to construct a set of candidate parking areas using the intent-intensive area, calculate the parking difficulty index of the set of candidate parking areas based on road segment complexity, historical occupancy rate, and playback simulation results, and simultaneously calculate the walking distance index from the set of candidate parking areas to the target area. The parking area recommendation list output module is used to read the user's parking preference information, perform parking adaptation filtering based on the parking difficulty index and walking distance index, and output a personalized parking area recommendation list.

Images