A map matching method and device based on trajectory topology

Abstract

The present invention discloses a map matching method and device based on trajectory topology, comprising the steps of S1: based on road network data, using three types of nodes, namely, entrances and exits, hubs, and connections, to model urban space and construct a city topology map; S2: using an original vehicle GPS record sequence to obtain a set of candidate points corresponding to the vehicle trajectory topology; S3: calculating a spatial similarity function between candidate points based on the original vehicle GPS record sequence and the city topology map to obtain a weighted candidate map; S4: establishing and calculating a static similarity matrix based on the weighted candidate map to obtain a weighted similarity matrix; S5: calculating a local optimal path, counting votes for all candidate points on the local optimal path, and connecting the candidate points with the highest number of votes to obtain a trajectory topology; and S6: performing map matching through the trajectory topology. The present invention uses the city topology map as a carrier for expressing trajectory data, simplifies the expression of the trajectory, and improves the accuracy and efficiency of matching the trajectory data with the map.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic big data, and in particular to a method and device for map matching based on trajectory topology. Background Art

[0002] Trajectory data is an emerging data source for intelligent transportation systems. It is derived from real-time location data reported by floating vehicles. Floating vehicles, or taxis equipped with a Global Positioning System (GPS), are an effective way to collect traffic data across large-scale road networks. These vehicles continuously upload their status information (such as latitude, longitude, instantaneous speed, and direction of movement) to data centers via wireless communication at short intervals, acting as ubiquitous mobile sensors monitoring urban traffic conditions.

[0003] Map matching is a key step in obtaining trajectory data from GPS data. Map matching aligns the observed position sequence with the road network on the map. Since urban space is continuous, the representation space of trajectory points is infinite. Therefore, it is necessary to topologically represent urban space. Previous research has commonly used methods to match trajectory points to road segments, then calculate traffic parameters on the observed road segments using the trajectory data, thereby observing traffic conditions. However, due to the low degree of topological representation of urban space, this approach results in a very time-consuming map matching process.

[0004] Existing solutions typically match GPS points to the entire road network, considering multiple possible paths through the road network to find the optimal matching path. However, due to the low degree of urban spatial topology, this method makes the map matching process quite time-consuming, and the representation of trajectory data is relatively complex. Summary of the Invention

[0005] The purpose of the present invention is to solve the problems of low map matching efficiency and high trajectory expression complexity in existing map matching methods, and to provide a map matching method and device based on trajectory topology.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A map matching method based on trajectory topology comprises the following steps: S1: based on road network data, urban space is modeled using three types of nodes: entrances and exits, hubs, and connections to construct an urban topology map; S2: using an original vehicle GPS record sequence to obtain a set of candidate points corresponding to the vehicle trajectory topology; S3: calculating a spatial similarity function between candidate points based on the original vehicle GPS record sequence and the urban topology map to obtain a weighted candidate map; S4: establishing and calculating a static similarity matrix based on the weighted candidate map to obtain a weighted similarity matrix; S5: calculating a local optimal path, counting votes for all candidate points on the local optimal path, and connecting the candidate points with the highest number of votes to obtain a trajectory topology; and S6: performing map matching using the trajectory topology.

[0008] In some embodiments of the present invention, step S1 includes the following steps: S11: obtaining road network data of the target area; S12: obtaining the entrance and exit coordinate data of the target area as the entrance and exit nodes; S13: obtaining the coordinate data of the road intersections in the target area based on the road network data as the hub nodes; S14: obtaining the coordinate data of the intersections of the main roads and branches in the target area based on the road network data, as well as the projection points of the entrance and exit nodes on the road network data as the connection nodes; S15: performing directed connections on the entrance and exit nodes, hub nodes, and connection nodes according to the road direction to construct a city topology map.

[0009] In some embodiments of the present invention, step S1 further includes step S16: establishing a spatial index for the nodes in the target area based on the R-Tree, so as to enable fast retrieval of all nodes in the target area.

[0010] In some embodiments of the present invention, in step S5, for all candidate points of each GPS point, it is assumed that the candidate point is a correct matching point of its corresponding GPS point, and a dynamic programming algorithm is used to obtain a local optimal path passing through the candidate point.

[0011] In some embodiments of the present invention, for each GPS point, based on the weighted similarity matrix, the dynamic programming algorithm is used twice to find the local optimal path with the candidate point of each GPS point as the starting point and the end point, and a vote is recorded once for each point on the local optimal path that is not a candidate point.

[0012] In some embodiments of the present invention, the steps of counting a vote for a point on the local optimal path that is not a candidate point are as follows:

[0013] S51: Find the optimal forward path, let dp_pre[t][j] represent the path from the first GPS point to the candidate point The maximum weighted similarity sum, the forward state transfer equation is:

[0014]

[0015] Where (1≤k≤a t-1 ) means traversing GPS points All candidate points of is the weighted similarity matrix Φ i Selected candidate points to The weighted similarity of , and the pre array is used to record the forward local optimal path as:

[0016]

[0017] S52: Find the optimal backward path, let dp_next[t][j] represent the candidate point is the maximum weighted similarity from the starting point to the last GPS point, and the backward state transfer equation is:

[0018]

[0019] is the weighted similarity matrix Φ i Selected candidate points to The weighted similarity is calculated, and the next array is used to record the backward local optimal path:

[0020]

[0021] S53: For each candidate point, use the pre array to find the passing candidate point The local optimal path of the last point is the best path until the first point, and the next array is used to find the next point of the optimal path until the last point. A vote is recorded for all non-candidate points found in the forward and backward directions.

[0022] In some embodiments of the present invention, in step S3, a spatial similarity function between the candidate points is calculated through position analysis and topological relationship analysis, and the spatial similarity function is:

[0023]

[0024] in, and is the candidate point corresponding to the two GPS points before and after, ∈ is the error factor, From the candidate point Move to another candidate point The probability function of .

[0025] In some embodiments of the present invention, in the position analysis, it is assumed that the GPS drift error conforms to the normal distribution N(μ,σ2 ), the error factor of the GPS point is:

[0026]

[0027] in, is the candidate point, It is a candidate point To GPS point P i Euclidean distance, μ is the average drift distance of GPS coordinates relative to the accurate position, σ 2 is the variance.

[0028] In some embodiments of the present invention, in step S4, the static similarity matrix is ​​calculated by introducing a distance decay function, and the expression of the distance decay function is:

[0029]

[0030] Where x is the distance between two GPS points and β is a constant.

[0031] The present invention also provides a map matching device based on trajectory topology, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the map matching method based on trajectory topology as described above is implemented.

[0032] The present invention also provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the above-mentioned trajectory topology-based map matching method.

[0033] The present invention has the following beneficial effects:

[0034] The map matching method based on trajectory topology proposed in the present invention models the urban space based on road network data using three types of nodes: entrances and exits, hubs, and connections, constructs an urban topology map, uses the original vehicle GPS record sequence to obtain a set of candidate points corresponding to the vehicle trajectory topology, and calculates the spatial similarity function and weighted similarity matrix between the candidate points, calculates the local optimal path, and obtains the trajectory topology through the interactive voting algorithm and other technical features. The method can use the urban topology map as the carrier for expressing trajectory data, simplifies the expression of trajectories compared to traditional trajectory data, and can improve the accuracy and efficiency of matching trajectory data with maps by using the interactive voting method.

[0035] In addition, some embodiments of the present invention also have the following beneficial effects:

[0036] By establishing a spatial index for the nodes in the target area based on R-Tree, all nodes in the target area can be quickly retrieved, thereby improving retrieval efficiency.

[0037] By using the dynamic programming algorithm twice in the interactive voting stage to find the optimal path, it is possible to perform only two calculations for each candidate set, without having to search for the local optimal path for each candidate point in turn, thereby improving operating efficiency, shortening growth time, and effectively saving costs.

[0038] Other beneficial effects of the embodiments of the present invention will be further described below. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flowchart of steps of an embodiment of the present invention;

[0040] Figure 2 is the GPS point P in Example 1 i Schematic diagram of candidate set generation;

[0041] Figure 3 is the weighted candidate graph between candidate points in Example 1;

[0042] Figure 4 It is the weight candidate graph between candidate points in Example 2. DETAILED DESCRIPTION

[0043] The present invention will be further described below with reference to the accompanying drawings and in combination with preferred embodiments. It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0044] It should be noted that the directional terms such as left, right, up, down, top, and bottom in this embodiment are merely relative concepts, or are based on the normal use status of the product, and should not be considered as restrictive.

[0045] The following embodiments of the present invention provide a trajectory topology-based map matching method to address the low efficiency and high complexity of trajectory representation in existing map matching methods. Using an interactive voting algorithm, this method matches GPS data to a discretized city topology model, generating a trajectory topology that is simplified compared to traditional trajectory data.

[0046] The embodiment of the present invention proposes a map matching method based on trajectory topology, such as Figure 1As shown, the method includes the following steps: S1: based on road network data, the urban space is modeled using three types of nodes: entrances and exits, hubs, and connections to construct an urban topology map; S2: using the original vehicle GPS record sequence to obtain a set of candidate points corresponding to the vehicle trajectory topology; S3: according to the original vehicle GPS record sequence and the urban topology map, the spatial similarity function between the candidate points is calculated to obtain a weighted candidate map; S4: according to the weighted candidate map, a static similarity matrix is ​​established and calculated to obtain a weighted similarity matrix; S5: calculating the local optimal path, counting votes for all candidate points on the local optimal path, and connecting the candidate points with the highest number of votes to obtain a trajectory topology; S6: performing map matching through the trajectory topology.

[0047] An embodiment of the present invention also provides a map matching device based on trajectory topology, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the map matching method based on trajectory topology as described above is implemented.

[0048] An embodiment of the present invention further provides a storage medium, which includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute the map matching method based on trajectory topology as described above.

[0049] Example 1

[0050] This embodiment provides a map matching method based on trajectory topology, including: S1, using three types of nodes: entrances and exits, hubs, and connections to model the urban space at a meso-level to obtain an indexed urban topology map; S2, using a sequence of GPS records generated by a vehicle-mounted GPS device to obtain a set of candidate points corresponding to the vehicle trajectory topology; S3, using the original GPS records and the topology map to perform position analysis and topological relationship analysis on the set of candidate points, calculating a spatial similarity function between the candidate points, and obtaining a weighted candidate map; S4, establishing a static similarity matrix based on the weighted candidate map, processing the static similarity matrix using a distance decay function, and obtaining a weighted similarity matrix; S5, using all candidate points using a dynamic programming algorithm to find a local optimal path passing through the candidate point, counting votes for all candidate points passed along the path, and using the candidate point with the highest number of votes corresponding to each GPS point as the global optimal path. These candidate points are sequentially connected in a directionally manner to obtain a trajectory topology.

[0051] 1. Create an indexed city topology map

[0052] The urban topology map is a meso-level urban space modeling obtained by connecting three types of functional points: entrances and exits, connections, and hubs in a directed manner.

[0053] (1) Obtaining the road network data corresponding to the actual target area through OpenStreetMap (an open source map);

[0054] (2) Obtain the regional entrance and exit coordinate data of the research target area through the Amap open platform (lbs.amap.com) as the entrance and exit nodes;

[0055] (3) Using Arcgis to process the road network data, the coordinate data of all road intersections are obtained as hub nodes;

[0056] (4) Using Arcgis to process the road network data, the coordinate data of the intersections of all road trunks and branches, as well as the projection points of the entrance and exit nodes in step 2 on the road network, are obtained as connecting nodes;

[0057] (5) Based on the road network data obtained in step 1, the three types of nodes are connected in a directed manner according to the road direction to establish a city topology map;

[0058] (6) Based on R-Tree (a spatial index data structure), a spatial index is established for all nodes in the study area, so as to quickly retrieve all functional points within the specified coordinate range.

[0059] Compared to the directed road network G(V,E), the topological map modeled using these three types of function points offers a higher degree of discretization, resulting in a more refined representation of trajectories. Furthermore, function points themselves carry semantics (their own functions), allowing for semantic attributes to be added to trajectory points that match them.

[0060] 2. Candidate Set Generation

[0061] In this stage, for the original vehicle GPS coordinate sequence, for each GPS coordinate point, the target point that may be matched to the point is found in the city topology map. All target points constitute the candidate set of the coordinate point.

[0062] Assume the original GPS trajectory is:

[0063] P1→P2→…→P n

[0064] At any point P i For example, we use r as the radius to perform range query based on Euclidean distance, where the default value of r is 100 meters. When there are too few candidate points in the range, we can increase it adaptively to obtain point P. i The candidate function point set CP i (candidate points), such as Figure 2 As shown, Construct GPS coordinate point P i candidate set.

[0065] The trajectory topology generated in this embodiment is expressed as a directed connection of function points in sequence. Therefore, when searching for candidate points, there is no need to pay attention to the road section, and the function points that may be matched within the search range can be directly searched, which speeds up the search efficiency.

[0066] 3. Spatial Similarity Function Calculation

[0067] In this stage, the spatial similarity function between candidate points is calculated through position analysis and topological relationship analysis for the candidate point set.

[0068] (1) Location analysis

[0069] Assume that GPS drift error conforms to normal distribution N(μ,σ 2 ), where μ is the average drift distance of GPS coordinates relative to the accurate position (m), σ 2 is the variance, then for the candidate point Its about GPS point P i The error factor is:

[0070]

[0071] in It is a candidate point To GPS point P i The default value of μ is 5 meters, and the default value of σ is 10 meters.

[0072] (2) Topological relationship analysis

[0073] definition From the candidate point Move to another candidate point The probability function of is calculated as follows:

[0074]

[0075] Where, dist(P i ,P i-1 ) is the Euclidean distance between the two GPS points. The shortest path length between two candidate points in the topology map. Since the total number of functional points in the topology map is approximately 7,000, the Floyd algorithm can be used to preprocess the shortest path lengths between all pairs of points in the topology map before matching to improve the efficiency of the matching operation.

[0076] (3) Calculation of spatial similarity function

[0077] Through location analysis and topological relationship analysis, define the candidate points corresponding to the two GPS points before and after and The spatial similarity function of :

[0078]

[0079] 4. Weighted Similarity Matrix Generation

[0080] First, define the static similarity matrix Figure 3 Taking the weighted candidate graph model in as an example, a similarity matrix can be constructed based on the graph:

[0081] M=diag{M (2) ,M (3) ,…,M (n)}

[0082] in Its meaning is the spatial similarity function of all candidate point transfer paths between candidate point set i-1 and candidate point set i, where a i is the number of candidate points corresponding to the i-th GPS point. The similarity matrix M reflects the similarity between adjacent GPS points, but does not reflect the impact of GPS points on the global path. Therefore, by introducing the distance decay function, a weighted similarity matrix is ​​constructed to extend the influence between adjacent sampling points to the global path. First, n block diagonal matrices are used to describe the mutual influence between the front and back points as the distance decays. Specifically, the distance decay function matrix W is defined as i ,in:

[0083]

[0084] in It is GPS point P j Relative to P i The dynamic weight of is obtained by calculating the distance decay function, where f is the distance decay function. Considering that the function reflects the characteristic that the influence between two GPS points decays as the distance increases, a negative exponential function is used as the distance decay function:

[0085]

[0086] Where x is the distance between two GPS points, and β is a fixed value, which is 7000 meters by default. From this, we can calculate the weighted similarity matrix Φ:

[0087] Φ i =W i M

[0088] The weighted similarity matrix reflects the GPS point P to be matched i , the influence of the matching of the points before and after it on the point, and the degree of influence decays with distance.

[0089] 5. Interactive Voting

[0090] In this phase, all candidate points for each GPS point are assumed to be the correct matching points for the corresponding candidate point. A dynamic programming algorithm is then used to find a local optimal path passing through the candidate point. Each candidate point on the local optimal path is counted once. Finally, the candidate points with the highest cumulative votes in each candidate point set are connected in sequence to form a trajectory topology.

[0091] Specifically, this stage traverses each GPS point P i , according to the weighted similarity matrix Φ generated in Section 4 i , use two dynamic programming algorithms to find is the local optimal path between the starting point and the end point, and is the optimal path for all non To record a vote, follow these steps:

[0092] (1) Find the optimal forward path. Let: dp_pre[t][j] represents the path from the first GPS point to the candidate point. The maximum weighted similarity sum, the forward state transfer equation is:

[0093]

[0094] Where (1≤k≤a t-1 ), which means traversing GPS point P t-1 All candidate points of Meaning is in the weighted similarity matrix Φ i In the candidate point to The weighted similarity of .

[0095] At the same time, use the pre array to record the forward local optimal path:

[0096]

[0097] The pre array records the previous point of each candidate point in its local optimal path.

[0098] (2) Find the optimal backward path, let: dp_next[t][j] represents the candidate point is the maximum weighted similarity from the starting point to the last GPS point, and the backward state transfer equation is:

[0099]

[0100] is the weighted similarity matrix Φ i Selected candidate points to The weighted similarity is calculated, and the next array is used to record the backward local optimal path:

[0101]

[0102] The next array records the next point of each candidate point in its local optimal path.

[0103] (3) For each candidate point Keep using the pre array to find the The local optimal path of the node is from the previous point to the first point, and the next array is used to find the next point of the optimal path until the last point. A vote is recorded for all the points found in the forward and backward directions.

[0104] (4) After all the voting is completed, the candidate point with the highest number of votes corresponding to each GPS point is regarded as the global optimal path. The route obtained by connecting these candidate points in a directed manner is the trajectory topology.

[0105] Example 2

[0106] by Figure 4 For example, the local optimal path of each point in CP2 is currently calculated, and the similarity of each edge of the weighted similarity matrix Φ2 is marked on the graph.

[0107] (1) Calculate dp_pre and pre: Initialize dp_pre[1][1] = dp_pre[1][2] = 0, calculate

[0108]

[0109] pre[2][1]=1

[0110]

[0111] pre[2][1]=2

[0112] (2) Calculate dp_next and next: Initialize dp_next[3][1] = dp_next[3][2] = 0, calculate

[0113]

[0114] next[2][1]=1

[0115]

[0116] next[2][2]=1

[0117] (3) Counting votes: Call pre until the first point, and find pre[2][1]=1, that is, the previous point is Find the path for Record one vote; call next until the last point to find the path for Record one vote. A similar process is performed, and finally the calculation of CP2 is completed.

[0118] Experimental example

[0119] In step S2, when generating candidate sets, r was set to 150 meters. In step S3, when calculating the spatial similarity function (position analysis), μ was set to 30 meters, σ was set to 100 meters, and during topological relationship analysis, β was set to 7000 meters. Using these parameters, the matching accuracy achieved 74%, compared to 68% for the comparable method, ST-Matching.

[0120] The embodiments of the present invention have the following features:

[0121] 1. Based on public data and historical data, the embodiment of the present invention can effectively generate trajectory topology data that can be used for traffic observation by constructing a refined urban topology model and an interactive voting algorithm.

[0122] 2. The method of the embodiment of the present invention has higher accuracy than existing similar methods and is a very promising method for generating trajectory data.

[0123] 3. This embodiment of the present invention uses a city topology map as the carrier for expressing trajectory data, simplifying the expression of trajectories; improving the accuracy of map matching by using an interactive voting method; and improving the algorithm by taking advantage of the topological characteristics of the city topology map, thereby improving the algorithm's operating efficiency compared to the original algorithm.

[0124] 4. The growth equipment of the embodiment of the present invention is simple, the growth time is short, and the cost is effectively saved.

[0125] 5. The topological graph constructed in the embodiment of the present invention has a higher degree of discretization than the directed road network G(V, E), and the functional points have semantic characteristics.

[0126] 6. The embodiment of the present invention directly searches for functional points based on the R-tree index when searching for a set of candidate points, which can improve efficiency.

[0127] 7. The embodiment of the present invention uses the Floyd algorithm to pre-process the shortest path between function points when calculating the weighted matrix, which can reduce the amount of calculation during matching.

[0128] 8. The embodiment of the present invention uses a two-step dynamic programming algorithm to find the optimal path during the interactive voting phase. Only two calculations are required for each candidate set, and there is no need to search for a local optimal path for each candidate point in turn.

[0129] The above is a further detailed description of the present invention in conjunction with specific preferred embodiments, and the specific implementation of the present invention should not be considered to be limited to these descriptions. Those skilled in the art will recognize that several equivalent substitutions or obvious variations can be made without departing from the scope of the present invention, and that any equivalent performance or application should be considered to fall within the scope of protection of the present invention.

Claims

1. A map matching method based on trajectory topology, characterized in that: The steps include: S1: Based on road network data, the urban space is modeled using three types of nodes: entrances and exits, hubs, and connections to construct a city topology map; S2: Using the original vehicle GPS record sequence, obtain the candidate point set corresponding to the vehicle trajectory topology; S3: Calculating a spatial similarity function between candidate points based on the original vehicle GPS record sequence and the city topology map to obtain a weighted candidate map; S4: establishing and calculating a static similarity matrix based on the weighted candidate graph to obtain a weighted similarity matrix; S5: Calculate the local optimal path, count votes for all candidate points on the local optimal path, and connect the candidate points with the highest votes to obtain the trajectory topology; S6: Perform map matching using the trajectory topology.

2. The map matching method based on trajectory topology according to claim 1, characterized in that: Step S1 includes the following steps: S11: Acquire road network data of the target area; S12: Obtaining the entrance and exit coordinate data of the target area as the entrance and exit nodes; S13: Obtaining coordinate data of road intersections in the target area based on the road network data as hub nodes; S14: obtaining, based on the road network data, coordinate data of intersections of trunk lines and branch lines in the target area, and projection points of the entrance and exit nodes on the road network data as connection nodes; S15: Directly connect the entrance and exit nodes, hub nodes, and connection nodes according to the road direction to construct a city topology map.

3. The map matching method based on trajectory topology according to claim 2, characterized in that: Step S1 also includes step S16: establishing a spatial index for the nodes in the target area based on the R-Tree, so as to enable fast retrieval of all nodes in the target area.

4. The map matching method based on trajectory topology according to claim 1, characterized in that: In step S5, for all candidate points of each GPS point, it is assumed that the candidate point is a correct matching point of its corresponding GPS point, and a dynamic programming algorithm is used to obtain a local optimal path passing through the candidate point.

5. The map matching method based on trajectory topology according to claim 4, characterized in that: For each GPS point, based on the weighted similarity matrix, the dynamic programming algorithm is used twice to find the local optimal path with the candidate point of each GPS point as the starting point and the end point, and a vote is recorded once for each point on the local optimal path that is not a candidate point.

6. The map matching method based on trajectory topology according to claim 5, characterized in that: The steps for counting a vote for a non-candidate point on the local optimal path are as follows: S51: Find the optimal forward path, let dp_pre[t][j] represent the path from the first GPS point to the candidate point The maximum weighted similarity sum, the forward state transfer equation is: Where (1≤k≤a t-1 ) means traversing GPS point P t-1 All candidate points of is the weighted similarity matrix Φ i Selected candidate points to The weighted similarity of , and the pre array is used to record the forward local optimal path as: S52: Find the optimal backward path, let dp_next[t][j] represent the candidate point is the maximum weighted similarity from the starting point to the last GPS point, and the backward state transfer equation is: is the weighted similarity matrix Φ i Selected candidate points to The weighted similarity is calculated, and the next array is used to record the backward local optimal path: S53: For each candidate point, use the pre array to find the passing candidate point The local optimal path of the last point is the best path until the first point, and the next array is used to find the next point of the optimal path until the last point. A vote is recorded for all non-candidate points found in the forward and backward directions.

7. The map matching method based on trajectory topology according to claim 1, characterized in that: In step S3, the spatial similarity function between the candidate points is calculated through position analysis and topological relationship analysis. The spatial similarity function is: in, and is the candidate point corresponding to the two GPS points before and after, ∈ is the error factor, From the candidate point Move to another candidate point The probability function of .

8. The map matching method based on trajectory topology according to claim 7, characterized in that: In the position analysis, it is assumed that the GPS drift error conforms to the normal distribution N(μ,σ 2 ), the error factor of the GPS point is: in, is the candidate point, It is a candidate point To GPS point P i Euclidean distance, μ is the average drift distance of GPS coordinates relative to the accurate position, σ 2 is the variance.

9. The map matching method based on trajectory topology according to claim 1, characterized in that: In step S4, the static similarity matrix is ​​calculated by introducing a distance decay function, and the expression of the distance decay function is: Where x is the distance between two GPS points and β is a constant.

10. A map matching device based on trajectory topology, characterized in that: The system comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the map matching method based on trajectory topology according to any one of claims 1 to 9 is implemented.

11. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute the map matching method based on trajectory topology according to any one of claims 1 to 9.

Images