A method, device and computer storage medium for reconstructing urban vehicle trajectories

Abstract

The embodiment of the present invention discloses a method, device and computer storage medium for reconstructing urban vehicle trajectories, comprising: obtaining initial vehicle trajectory data, identifying an incomplete trajectory data set of two non-adjacent intersections between continuous intersections; based on the incomplete trajectory data set, obtaining k trajectories between the two non-adjacent intersections as initial particles; inputting the initial particles and a spatiotemporal correction factor into a particle filter, resampling the basic correction factor and the extended correction factor satisfying the roulette method, and obtaining a reconstructed trajectory; wherein the spatiotemporal correction factor includes a basic correction factor and an extended correction factor; in this way, by classifying the spatiotemporal correction factor and designing different resampling processes, the influence of subjective factors on the result is reduced while retaining the interpretability of the spatiotemporal correction factor, and the randomness of the reconstruction is reduced.

Description

Technical Field

[0001] The present invention relates to the field of traffic big data analysis, and in particular to a method and device for reconstructing urban vehicle trajectories and a computer storage medium. Background Art

[0002] Vehicle trajectory data has both temporal and spatial attributes and can provide data support for many studies, such as travel time estimation, vehicle emission estimation, and OD mode estimation. With the rapid development of location tracking and storage technology in recent years, people have been able to obtain a large amount of vehicle trajectory data, but due to the limitations of equipment conditions, there are still some incomplete trajectories in these trajectory data. Incomplete vehicle trajectories will greatly limit the application of trajectory data in subsequent research. Therefore, reconstructing vehicle trajectory data is of great significance for further mining of trajectory data.

[0003] At present, the trajectory reconstruction method can be roughly summarized into two stages: the first stage is the trajectory set generation stage, and the main algorithms include DFS algorithm, KSP algorithm, K-shortest path algorithm (P-KSP) based on space-time prism improvement, etc.; the trajectories searched by the algorithms in the first stage can basically cover the real trajectories, so the difference in algorithms in this stage often does not affect the accuracy of the reconstruction results. The second stage is the optimal trajectory decision stage, and the main algorithms include TOPSIS algorithm, GRA algorithm, particle filter algorithm, GBDT algorithm, autoencoder algorithm, etc. The algorithm in this stage needs to determine several decision features about the path (also called "space-time correction factor") at the beginning, and then substitute these attribute indicators into their respective models to obtain the optimal trajectory. Therefore, the key to the final accuracy of the algorithm lies in the determination of decision features and the selection of models. Except for the GBDT algorithm, most of the decision features determined by the above studies have certain subjective factors. The existence of subjective factors makes the decision features ignore the travel preferences of a few people and increases the randomness of reconstruction. However, the decision features determined by the GBDT algorithm have the disadvantages of poor interpretability and inconsistency between different regions. How to balance the relationship between the subjective factors and interpretability of decision features is still a difficult problem that needs to be solved at present. Summary of the invention

[0004] In view of this, the embodiments of the present invention provide a method, device and computer storage medium for reconstructing urban vehicle trajectories, which reduce the influence of subjective factors on the results while retaining the interpretability of the spatiotemporal correction factor, and reduce the randomness of the reconstruction.

[0005] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0006] In a first aspect, an embodiment of the present invention provides a method for reconstructing urban vehicle trajectories, the method comprising:

[0007] Obtaining initial vehicle trajectory data, identifying incomplete trajectory data sets of two non-adjacent intersections between consecutive intersections;

[0008] Based on the incomplete trajectory data set, k trajectories between the two non-adjacent intersections are obtained as initial particles;

[0009] The initial particles and the spatiotemporal correction factors are input into a particle filter, and the basic correction factors and the extended correction factors satisfying the roulette method are resampled to obtain a reconstructed trajectory; wherein the spatiotemporal correction factors include the basic correction factors and the extended correction factors.

[0010] The step of acquiring the initial trajectory data of the vehicle and identifying an incomplete trajectory data set of two non-adjacent intersections between consecutive intersections includes:

[0011] Acquire initial trajectory data of the vehicle, and obtain multiple intersections and their positions included in the initial trajectory data;

[0012] It is confirmed that the current intersection and the next intersection are not adjacent to each other, and an incomplete trajectory dataset is obtained.

[0013] The step of obtaining k trajectories between two non-adjacent intersections as initial particles based on the incomplete trajectory data set includes:

[0014] Based on the incomplete trajectory data set, obtaining a starting intersection and a key intersection corresponding to each of the incomplete trajectory data;

[0015] Through the k shortest path algorithm, k trajectories are obtained as initial particles.

[0016] The basic correction factors include travel time consistency factor and measurability criterion factor, and the extended correction factors include path size factor, number of turns factor and distance criterion factor.

[0017] The resampling of the basic correction factor and the extended correction factor satisfying the roulette method comprises:

[0018] The basic correction factor is resampled respectively, and the roulette method is performed on the extended correction factor. If the result of the roulette method is within a preset probability interval, the resampling of the extended correction factor is performed. If the result of the roulette method is not within the preset probability interval, it is not performed.

[0019] In a second aspect, an embodiment of the present invention provides a device for reconstructing urban vehicle trajectories, the device comprising:

[0020] A data acquisition module, used to acquire initial trajectory data of the vehicle and identify incomplete trajectory data sets of two non-adjacent intersections between consecutive intersections;

[0021] A data processing module, used for obtaining k trajectories between the two non-adjacent intersections as initial particles based on an incomplete trajectory data set;

[0022] The trajectory reconstruction module is used to input the initial particles and the spatiotemporal correction factors into the particle filter, resample the basic correction factors and the extended correction factors that satisfy the roulette method, and obtain the reconstructed trajectory; wherein the spatiotemporal correction factors include the basic correction factors and the extended correction factors.

[0023] In a third aspect, an embodiment of the present invention provides a device for reconstructing urban vehicle trajectories, the device comprising: a processor and a memory for storing a computer program that can be run on the processor,

[0024] Wherein, when the processor is used to run the computer program, it implements the urban vehicle trajectory reconstruction method described in the first aspect.

[0025] In a fourth aspect, an embodiment of the present invention provides a computer storage medium storing a computer program, wherein when the computer program is executed by a processor, the urban vehicle trajectory reconstruction method according to the first aspect is implemented.

[0026] An embodiment of the present invention provides a method, device and computer storage medium for reconstructing urban vehicle trajectories, comprising: obtaining initial vehicle trajectory data, identifying an incomplete trajectory data set of two non-adjacent intersections between continuous intersections; based on the incomplete trajectory data set, obtaining k trajectories between the two non-adjacent intersections as initial particles; inputting the initial particles and the spatiotemporal correction factor into a particle filter, resampling the basic correction factor and the extended correction factor that satisfies the roulette method, and obtaining a reconstructed trajectory; wherein the spatiotemporal correction factor includes a basic correction factor and an extended correction factor; in this way, by classifying the spatiotemporal correction factor and designing different resampling processes, the influence of subjective factors on the results is reduced while retaining the interpretability of the spatiotemporal correction factor, thereby reducing the randomness of the reconstruction. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A schematic diagram of a flow chart of a method for reconstructing urban vehicle trajectories provided by an embodiment of the present invention;

[0028] Figure 2 A schematic diagram of a flow chart of a method for reconstructing urban vehicle trajectories provided by another embodiment of the present invention;

[0029] Figure 3A schematic diagram of a particle filter model flow chart provided by an embodiment of the present invention;

[0030] Figure 4 A schematic diagram of the structure of a device for reconstructing urban vehicle trajectories provided by an embodiment of the present invention;

[0031] Figure 5 A schematic diagram of the structure of another urban vehicle trajectory reconstruction device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The technical solution of the present invention is further elaborated in detail below in conjunction with the accompanying drawings and specific embodiments of the specification. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art of the technical field of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0033] See also Figure 1 , a method for reconstructing urban vehicle trajectories provided by an embodiment of the present invention, the method for reconstructing urban vehicle trajectories can be applied to the situation of predicting the impact range of traffic accidents on road sections, the method for reconstructing urban vehicle trajectories can be executed by a device for reconstructing urban vehicle trajectories provided by an embodiment of the present invention, the device for reconstructing urban vehicle trajectories can be implemented in software and / or hardware, in a specific application, the device for reconstructing urban vehicle trajectories can be a terminal such as a desktop computer, a laptop computer, a smart phone, a personal digital assistant, a tablet computer, etc. The method for reconstructing urban vehicle trajectories includes the following steps:

[0034] Step S101: obtaining initial vehicle trajectory data, and identifying incomplete trajectory data sets of two non-adjacent intersections between consecutive intersections;

[0035] Here, the initial vehicle trajectory data is usually collected by two types of sensors: mobile sensors and fixed sensors. Mobile sensor data includes GPS data, smartphone data, and smart card data; fixed sensor data includes geomagnetic detection data and automatic license plate recognition (ALPR) data. The initial vehicle trajectory data here mainly refers to ALPR data.

[0036] Step S102: Based on the incomplete trajectory data set, k trajectories between the two non-adjacent intersections are obtained as initial particles;

[0037] Here, the initial vehicle trajectory data can be divided into two categories: one is a complete travel trajectory, that is, the intersections passed in sequence in the trajectory are adjacent; the other is an incomplete travel trajectory, that is, there are breakpoints in the trajectory, and there are two non-adjacent intersections between consecutive intersections. The main function of this part of the model is to find the incomplete trajectory and the position that needs to be reconstructed in the incomplete trajectory.

[0038] Step S103: inputting the initial particles and the spatiotemporal correction factors into a particle filter, resampling the basic correction factors and the extended correction factors satisfying the roulette method, and obtaining a reconstructed trajectory; wherein the spatiotemporal correction factors include the basic correction factors and the extended correction factors.

[0039] In the above embodiment of the present invention, the initial trajectory data of the vehicle is obtained, and an incomplete trajectory data set of two non-adjacent intersections between continuous intersections is identified; based on the incomplete trajectory data set, k trajectories between the two non-adjacent intersections are obtained as initial particles; the initial particles and the spatiotemporal correction factors are input into a particle filter, and the basic correction factors and the extended correction factors that satisfy the roulette method are resampled to obtain a reconstructed trajectory; wherein the spatiotemporal correction factors include basic correction factors and extended correction factors; in this way, by classifying the spatiotemporal correction factors and designing different resampling processes, the influence of subjective factors on the results is reduced while retaining the interpretability of the spatiotemporal correction factors, and the randomness of the reconstruction is reduced.

[0040] In one embodiment, initial trajectory data of the vehicle is acquired to obtain a plurality of intersections and their positions included in the initial trajectory data;

[0041] It is confirmed that the current intersection and the next intersection are not adjacent to each other, and an incomplete trajectory dataset is obtained.

[0042] Here, please see Figure 2 , the current intersection and the next intersection are not adjacent, that is, it is confirmed that there is a breakpoint in the trajectory. The algorithm steps for breakpoint identification are as follows:

[0043] Initialize, set k max initial trajectories, each containing h max intersection, let the trajectory number k = 0, the trajectory intersection number h = 0;

[0044] Step 1: k = k + 1, if k = k max , all trajectories are retrieved, otherwise go to step 2;

[0045] Step 2: h = h + 1, if h = h max , all intersections of the trajectory have been retrieved, go to step 1, otherwise go to step 3;

[0046] Step 3: Determine whether the current intersection is adjacent to the next intersection. If so, go to step 2. Otherwise, store the current intersection s and the next intersection t in the breakpoint matrix and go to step 3.

[0047] In one embodiment, obtaining k trajectories between the two non-adjacent intersections as initial particles based on the incomplete trajectory data set includes:

[0048] Based on the incomplete trajectory data set, obtaining a starting intersection and a key intersection corresponding to each of the incomplete trajectory data;

[0049] Through the k shortest path algorithm, k trajectories are obtained as initial particles.

[0050] Here, in order to find the k shortest paths between the starting intersection and the key intersection, the present invention directly quotes the MPS algorithm proposed by Professor Martins in 1998. The parameters required for input of the algorithm are the impedance matrix M and the road network G = (N, A), where N is the set of nodes in the road network and A is the set of road sections in the road network.

[0051] The specific steps of the algorithm are as follows:

[0052] Step 1: Use Dijkstra algorithm to find the shortest path from the starting point s to the end point t in the directed graph, denoted as p k , at this time k=1;

[0053] Step 2: If k is less than the maximum number K of the required shortest paths, and there are still candidate paths, go to 3, otherwise, the algorithm ends;

[0054] Step 3: Find out the current path p k The first node with in-degree greater than 1 starting from the first node is denoted as n h If n h There is no extended node n′ h , go to 4. If it exists, find the path p k Medium h The first node that does not have an expansion node is recorded as n i , turn 5;

[0055] Step 4: For node n h Construct an extended node n′n and add it to the set N, and at the same time select all n from the graph (N, A) h The predecessor node is connected to n′ n The weights of the arcs remain unchanged, and these arcs are added to the arc set A, but n h In the path p k Except for the previous node in. And record n i =n h+1, turn 5;

[0056] Step 5: For p k From n i All subsequent nodes that begin with this command will perform the following operations in sequence:

[0057] 1) Add n i The extended node n′ i to the node set N.

[0058] 2) In addition to path p k Medium i In addition to the previous node, each node is connected to a node from n i Predecessor node to its extended node n i The weights of the arcs remain unchanged, and these arcs are added to the arc set A. In addition, if p k Medium i The previous node (denoted as n i-1 ) has an extended node (denoted as n′ i-1 ), we also need to connect a line from n′ i-1 to n i The arc, weight and arc (n i-1 , n i ) have equal weights.

[0059] Step 6: Use Dijkstra algorithm to obtain the current graph G = (N, A) from the start node s to the end node of the current expansion node t' k The shortest path between them is the kth shortest path, let k=k+1, and go to step 2.

[0060] In one embodiment, the basic correction factor includes a travel time consistency factor and a measurability criterion factor, and the extended correction factor includes a path size factor, a turn number factor, and a distance criterion factor.

[0061] Here, the travel time consistency factor and measurability criterion factor are proposed based on objective facts and are called basic correction factors; while the path size factor, number of turns factor, and distance criterion factor are proposed based on subjective factors and are called extended correction factors.

[0062] In one embodiment, resampling the base correction factor and the extended correction factor satisfying the roulette method comprises:

[0063] The basic correction factor is resampled respectively, and the roulette method is performed on the extended correction factor. If the result of the roulette method is within a preset probability interval, the resampling of the extended correction factor is performed. If the result of the roulette method is not within the preset probability interval, it is not performed.

[0064] Here, five important sampling processes are based on five correction factors. The resampling process of the extended correction factor needs to be performed once before it starts. For example, the probability α of the roulette method is set in advance. If the result of the roulette method is within the selected probability interval, then all particles will execute the resampling process of the extended correction factor; otherwise, it will not be executed. The resampling process of the basic correction factor does not need to be performed before it starts, and the following content can be directly executed.

[0065] Please refer again Figure 2 and Figure 3 , the particle filter model process in the urban vehicle trajectory reconstruction method is described through a specific embodiment, as follows.

[0066] Step 1: Initialize the initial particles (x 1 , x 2 , ..., x I );

[0067] The trajectory x between two non-adjacent intersections found 1 , x 2 , ..., xI as the initial particle, P(x 1 ), P(x 2 ), .., P(x I ) represents the prior probability of the initial particle. In the absence of historical trajectory data, the prior probability of a particle is set to 1 / I, where I represents the number of particles.

[0068] Step 2: Importance Sampling

[0069] For i=1, 2, ..., I, it is assumed that the prior distribution of the particles obeys the density function obtained by the previous importance sampling. The five important sampling processes are based on five correction factors. The resampling process of the extended correction factor needs to be performed once before starting the roulette method, and the probability α of the roulette method is set in advance. If the result of the roulette method is within the selected probability interval, then all particles will execute the resampling process of the extended correction factor; otherwise, it will not be executed. The resampling process of the basic correction factor does not need to be performed before starting the roulette method, and the following content can be directly executed.

[0070] In the first importance sampling (travel time consistency), the initial particle mass follows a uniform distribution and can be expressed as represents the travel time consistency probability density function, and the posterior probability distribution obeys

[0071] The weight update equation for travel time consistency is:

[0072]

[0073] in, represents the unnormalized weight of possible trajectory i after travel time consistency update.

[0074] Represents the initial prior weight of possible trajectory i, which is 1 at the beginning.

[0075] y t：t+Δt represents the objective data or empirical standard collected from time t to time t+Δt, where t is the time when data collection starts and Δt is a variable that depends only on the vehicle trajectory reconstruction calculation interval.

[0076] represents the state space of possible trajectories i.

[0077] Indicates that the initial possible trajectory is particle i.

[0078] represents the probability of a possible trajectory i based on the consistency of travel time.

[0079] represents the transition probability from prior data to travel time consistency. For a static calculation process, there is no spatial or temporal transition relationship between possible trajectories. Therefore, the state transition probability can be regarded as a constant.

[0080] Represents the prior probability density function, which can be regarded as a constant probability density function.

[0081] In order to simplify the calculation process, we do not consider the normalization calculation of particle weights; instead, we use the particle aggregation equation for calculation, where the particle weights are the updated particle weights. The particle aggregation equation is expressed as:

[0082]

[0083] in Represents the number of particle clusters for possible trajectory i after travel time consistency calculation.

[0084] After the first importance sampling, there is also a resampling of the base correction factor and a resampling of several expansion correction factors. The first importance sampling process is called step 2.1, and the calculation process is as follows: step (1) - step (2.1) - step (2.2) - step (2.3) / (2.4) / (2.5) - step (3).

[0085] Step 3: Output the real trajectory

[0086] After several resampling processes, the possible trajectory with the largest number of particle aggregations is output as the optimal trajectory.

[0087] The above is the entire process of improving the particle filter model. Next, we will further introduce the spatiotemporal correction factors.

[0088] 3.1 The first correction factor: travel time consistency

[0089] In previous studies, travel time consistency is described as follows: the average travel time of the path selected by the vehicle is close to the actual travel time between the two AVI devices. Under this description, travel time consistency is a group calculation process, as shown in formula (3), the average travel time of the possible paths between the two bayonet devices is compared with the actual travel time, and the path that is closest to the two is often the path selected by the vehicle. In the study, the present invention proposes a new understanding of travel time consistency: compared with the path not selected by the vehicle, the actual travel time has a higher probability in the travel time distribution of the path selected by the vehicle. Based on this understanding, we need to first fit the travel time distribution function of all possible paths (the present invention uses a Gaussian mixture model), and then substitute the actual travel time into each distribution function to obtain the result, as shown in formula (4).

[0090] Traditional travel time consistency compares the actual travel time between two checkpoints with the average travel time using the following formula:

[0091]

[0092] in, represents the posterior probability of possible trajectory i based on travel time consistency;

[0093] represents the average travel time of possible trajectory i between two AVI devices;

[0094] t'(AVI1, AVI2) represents the actual travel time between two AVI devices.

[0095] The improved travel time consistency first fits the travel time distribution function of all possible paths, and the posterior probability of possible trajectory i is also determined by this distribution function:

[0096]

[0097] Among them, F i (t) represents the travel time distribution function of possible trajectory i based on the Gaussian mixture model;

[0098] Initial weights for all trajectories are equal and are considered as 1 / N, where N is the number of trajectories. Follow formula (4). The probability density function represents the consistency of travel time. Since the consistency of travel time is an objective feature, this function can be regarded as a constant probability density function.

[0099] 3.2 The second correction factor: measurability criterion

[0100] The testability criterion is a reverse reasoning process based on the detection error of the AVI device. There are two reasons why the AVI device may not detect the vehicle: the first reason is that the vehicle passed the intersection where the AVI device was set up, but the vehicle was not detected due to the device detection error; the other reason is that the vehicle did not pass the intersection where the AVI device was set up. Based on these two reasons, when the possible path i between two AVI devices contains other intersections or links where AVI devices are set up, we can assume that the AVI devices at these intersections or links have made errors in detection.

[0101] therefore, It should obey the probability density function of formula (5), It can be viewed as a constant probability density function.

[0102]

[0103] in, represents the posterior probability of possible trajectory i based on the measurability criterion;

[0104] ε represents the probability of AVI device detection error;

[0105] α represents the number of possible intersections containing other AVI devices in trajectory i.

[0106] 3.3 The third correction factor: Path-Size

[0107] The concept of path-size is proposed in the path-size logit model, which is mainly used to solve the path overlap problem in path selection behavior. In this model, a deterministic correction term called path size (PS) is added to the utility function. PS is defined as follows (6):

[0108]

[0109] Among them, PS in represents the size of the path size of possible trajectory i;

[0110] Γ i represents the set of all possible segments of trajectory i;

[0111] l k represents the length of road segment k;

[0112] L i represents the length of possible trajectory i;

[0113] When segment k is on trajectory j, δ kj The value of is 1; when the road segment k is not on the trajectory j, δ ki The value of is 0.

[0114] If the trajectory i and the trajectory set C n The other paths in do not overlap, so their PS in The value is 1. Otherwise, the more segments it shares with other paths, the higher its PS in The lower the value.

[0115] The present invention uses the path size as a correction factor, when the PS of possible trajectory i in The lower the value, the higher its heat, so the more vehicles choose this trajectory, the more likely it is to become the real path. The results described in Section 3.2 are selected as the prior weights. Let The probability density function in the formula is obeyed. It can be regarded as a constant probability density function as follows (7).

[0116]

[0117] in, represents the posterior probability of possible trajectory i based on the path size criterion.

[0118] PS in represents the size of the path size of possible trajectory i;

[0119] M represents the set of all possible trajectories.

[0120] 3.4 The fourth correction factor: number of turns

[0121] Through observation of real vehicle trajectories, we found that in general, drivers have clear destinations and will choose a more appropriate path based on the destination. They rarely make frequent turns, lane changes, U-turns, detours, etc. Therefore, we describe the correction factor as follows: If the number of turns of the possible trajectory i of the vehicle is at a low level in the candidate trajectory set, then the trajectory is more likely to become the real trajectory. Select the result described in 3.3 as the prior weight. Let The probability density function in the obeying formula is, It can be regarded as a constant probability density function as follows (8).

[0122]

[0123] in, represents the posterior probability of possible trajectory i based on the number of turns criterion;

[0124] N i represents the number of turns of possible trajectory i;

[0125] M represents the set of all possible trajectories.

[0126] 3.5 Fifth correction factor: driving distance

[0127] Since drivers usually give priority to paths with shorter distances when choosing paths, the travel distance correction factor can be described as follows: if the distance of the vehicle's possible trajectory i is at a low level in the candidate trajectory set, then the trajectory is more likely to become the true trajectory. The results described in Section 3.4 are selected as the prior weights. Let The probability density function in the obeying formula is, It can be regarded as a constant probability density function as follows (9).

[0128]

[0129] in, represents the posterior probability of possible trajectory i based on the number of turns criterion;

[0130] D i represents the driving distance of possible trajectory i;

[0131] M represents the set of all possible trajectories.

[0132] Based on the same inventive concept as the above embodiments, see Figure 4 , which shows the composition of a device for reconstructing urban vehicle trajectories provided by an embodiment of the present invention, which may include: a data acquisition module 10, a data processing module 20, and a trajectory reconstruction module 30; wherein,

[0133] The data acquisition module 10 is used to acquire the initial trajectory data of the vehicle and identify the incomplete trajectory data set of two non-adjacent intersections between the continuous intersections;

[0134] A data processing module 20 is used to obtain k trajectories between the two non-adjacent intersections as initial particles based on the incomplete trajectory data set;

[0135] The trajectory reconstruction module 30 is used to input the initial particles and the spatiotemporal correction factors into a particle filter, resample the basic correction factors and the extended correction factors satisfying the roulette method, and obtain a reconstructed trajectory; wherein the spatiotemporal correction factors include basic correction factors and extended correction factors.

[0136] Optionally, the data acquisition module 10 is further used to acquire initial trajectory data of the vehicle, and obtain multiple intersections and their positions included in the initial trajectory data;

[0137] It is confirmed that the current intersection and the next intersection are not adjacent to each other, and an incomplete trajectory dataset is obtained.

[0138] Optionally, the data processing module 20 is further configured to obtain a starting intersection and a key intersection corresponding to each of the incomplete trajectory data based on the incomplete trajectory data set;

[0139] Through the k shortest path algorithm, k trajectories are obtained as initial particles.

[0140] Optionally, the trajectory reconstruction module 30 is also used to resample the basic correction factor and perform a roulette method on the extended correction factor respectively. If the result of the roulette method is within a preset probability interval, resampling of the extended correction factor is performed; if the result of the roulette method is not within the preset probability interval, resampling of the extended correction factor is not performed.

[0141] In another embodiment, Figure 5 As shown, a computer device is also provided, comprising: at least one processor 210 and a memory 211 for storing a computer program that can be run on the processor 210; wherein, Figure 5 The processor 210 shown in the figure is not used to indicate that the number of processors is one, but is only used to indicate the position relationship of the processor relative to other devices. In actual applications, the number of processors may be one or more; similarly, Figure 5 The memory 211 shown in the figure has the same meaning, that is, it is only used to refer to the position relationship of the memory relative to other devices. In actual applications, the number of memories can be one or more.

[0142] The processor 210 is used to execute the following steps when running the computer program:

[0143] Obtaining initial vehicle trajectory data, identifying incomplete trajectory data sets of two non-adjacent intersections between consecutive intersections;

[0144] Based on the incomplete trajectory data set, k trajectories between the two non-adjacent intersections are obtained as initial particles;

[0145] The initial particles and the spatiotemporal correction factors are input into a particle filter, and the basic correction factors and the extended correction factors satisfying the roulette method are resampled to obtain a reconstructed trajectory; wherein the spatiotemporal correction factors include the basic correction factors and the extended correction factors.

[0146] In an optional embodiment, the processor 210 is further configured to execute the following steps when running the computer program:

[0147] Acquire initial trajectory data of the vehicle, and obtain multiple intersections and their positions included in the initial trajectory data;

[0148] It is confirmed that the current intersection and the next intersection are not adjacent to each other, and an incomplete trajectory dataset is obtained.

[0149] In an optional embodiment, the processor 210 is further configured to execute the following steps when running the computer program:

[0150] Based on the incomplete trajectory data set, obtaining a starting intersection and a key intersection corresponding to each of the incomplete trajectory data;

[0151] Through the k shortest path algorithm, k trajectories are obtained as initial particles.

[0152] In an optional embodiment, the processor 210 is further configured to execute the following steps when running the computer program:

[0153] The basic correction factors include a travel time consistency factor and a measurability criterion factor, and the extended correction factors include a path size factor, a turn number factor, and a distance criterion factor.

[0154] In an optional embodiment, the processor 210 is further configured to execute the following steps when running the computer program:

[0155] The basic correction factor is resampled respectively, and the roulette method is performed on the extended correction factor. If the result of the roulette method is within a preset probability interval, the resampling of the extended correction factor is performed. If the result of the roulette method is not within the preset probability interval, it is not performed.

[0156] The computer device may also include: at least one network interface 212. The various components in the sending end are coupled together through a bus system 213. It is understood that the bus system 213 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 213 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, Figure 5 Various buses are labeled as bus system 213.

[0157] The memory 211 may be a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memories. The nonvolatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a ferromagnetic random access memory, a flash memory, a magnetic surface memory, an optical disk, or a compact disc read-only memory (CD-ROM); the magnetic surface memory may be a disk memory or a tape memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAMbus random access memory (DRRAM).The memory 211 described in the embodiments of the present invention is intended to include but is not limited to these and any other suitable types of memories.

[0158] The memory 211 in the embodiment of the present invention is used to store various types of data to support the operation of the sending end. Examples of these data include: any computer program used to operate on the sending end, such as an operating system and an application program. Among them, the operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application program can include various application programs for implementing various application services. Here, the program for implementing the method of the embodiment of the present invention can be included in the application program.

[0159] This embodiment also provides a computer storage medium, for example, including a memory 211 storing a computer program, and the computer program can be executed by the processor 210 in the sending end to complete the steps of the above method. The computer storage medium can be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disk, or CD-ROM; it can also be various devices including one or any combination of the above memories, such as smart phones, tablet computers, laptops, etc. A computer storage medium, wherein the computer storage medium stores a computer program, and when the computer program is executed by the processor, the following steps are performed:

[0160] The processor 210 is used to execute the following steps when running the computer program:

[0161] Obtaining initial vehicle trajectory data, identifying incomplete trajectory data sets of two non-adjacent intersections between consecutive intersections;

[0162] Based on the incomplete trajectory data set, k trajectories between the two non-adjacent intersections are obtained as initial particles;

[0163] The initial particles and the spatiotemporal correction factors are input into a particle filter, and the basic correction factors and the extended correction factors satisfying the roulette method are resampled to obtain a reconstructed trajectory; wherein the spatiotemporal correction factors include the basic correction factors and the extended correction factors.

[0164] In an optional embodiment, when the computer program is executed by a processor, the following steps are further performed:

[0165] Acquire initial trajectory data of the vehicle, and obtain multiple intersections and their positions included in the initial trajectory data;

[0166] It is confirmed that the current intersection and the next intersection are not adjacent to each other, and an incomplete trajectory dataset is obtained.

[0167] In an optional embodiment, when the computer program is executed by a processor, the following steps are further performed:

[0168] Based on the incomplete trajectory data set, obtaining a starting intersection and a key intersection corresponding to each of the incomplete trajectory data;

[0169] Through the k shortest path algorithm, k trajectories are obtained as initial particles.

[0170] In an optional embodiment, when the computer program is executed by a processor, the following steps are further performed:

[0171] The basic correction factors include a travel time consistency factor and a measurability criterion factor, and the extended correction factors include a path size factor, a turn number factor, and a distance criterion factor.

[0172] In an optional embodiment, when the computer program is executed by a processor, the following steps are further performed:

[0173] The basic correction factor is resampled respectively, and the roulette method is performed on the extended correction factor. If the result of the roulette method is within a preset probability interval, the resampling of the extended correction factor is performed. If the result of the roulette method is not within the preset probability interval, it is not performed.

[0174] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0175] In this document, the terms "comprises," "comprising," or any other variations thereof, are intended to cover a non-exclusive inclusion of elements other than those listed and may also include additional elements not expressly listed.

[0176] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for reconstructing urban vehicle trajectories, characterized in that: The method comprises: Obtaining initial vehicle trajectory data, identifying incomplete trajectory data sets of two non-adjacent intersections between consecutive intersections; Based on the incomplete trajectory data set, k trajectories between the two non-adjacent intersections are obtained as initial particles; Inputting the initial particles and the spatiotemporal correction factors into a particle filter, resampling the basic correction factors and the extended correction factors satisfying the roulette method, and obtaining a reconstructed trajectory; wherein the spatiotemporal correction factors include the basic correction factors and the extended correction factors; The step of obtaining k trajectories between the two non-adjacent intersections as initial particles based on the incomplete trajectory data set includes: Based on the incomplete trajectory data set, obtaining a starting intersection and a key intersection corresponding to each of the incomplete trajectory data; Through the k shortest path algorithm, k trajectories are obtained as initial particles; The basic correction factors include travel time consistency factor and measurability criterion factor, and the extended correction factors include path size factor, number of turns factor, and distance criterion factor; The resampling of the basic correction factor and the extended correction factor satisfying the roulette method comprises: The basic correction factor is resampled respectively, and the roulette method is performed on the extended correction factor. If the result of the roulette method is within a preset probability interval, the resampling of the extended correction factor is performed; if the result of the roulette method is not within the preset probability interval, the resampling of the extended correction factor is not performed.

2. The urban vehicle trajectory reconstruction method according to claim 1, characterized in that: The obtaining of the initial trajectory data of the vehicle and identifying an incomplete trajectory data set of two non-adjacent intersections between the continuous intersections includes: Acquire initial trajectory data of the vehicle, and obtain multiple intersections and their positions included in the initial trajectory data; It is confirmed that the current intersection and the next intersection are not adjacent to each other, and an incomplete trajectory dataset is obtained.

3. A device for reconstructing urban vehicle trajectories, characterized in that: The device comprises: A data acquisition module, used to acquire initial trajectory data of the vehicle and identify incomplete trajectory data sets of two non-adjacent intersections between consecutive intersections; The data processing module is used to obtain k trajectories between the two non-adjacent intersections as initial particles based on the incomplete trajectory data set; and is also used to obtain the corresponding starting intersection and key intersection in each incomplete trajectory data based on the incomplete trajectory data set; and obtain k trajectories as initial particles through a k-shortest path algorithm; The trajectory reconstruction module is used to input the initial particles and the spatiotemporal correction factors into the particle filter, resample the basic correction factors and the extended correction factors that satisfy the roulette method, and obtain the reconstructed trajectory; it is also used to resample the basic correction factors and perform the roulette method on the extended correction factors respectively, if the result of the roulette method is within the preset probability interval, then resampling of the extended correction factors is performed, if the result of the roulette method is not within the preset probability interval, then it is not performed; wherein, the spatiotemporal correction factors include basic correction factors and extended correction factors; the basic correction factors include travel time consistency factors and measurability criterion factors, and the extended correction factors include path size factors, number of turns factors, and distance criterion factors.

4. A device for reconstructing urban vehicle trajectories, characterized in that: include: a processor and a memory for storing a computer program capable of being executed on the processor; Wherein, when the processor is used to run the computer program, it implements the urban vehicle trajectory reconstruction method described in any one of claims 1 to 2.

5. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the urban vehicle trajectory reconstruction method according to any one of claims 1 to 2 is implemented.

Images