A method and system for stitching vehicle trajectories throughout a highway

Abstract

A method and system for stitching vehicle trajectories throughout a highway, which relates to the field of intelligent transportation. The up and down directions of the highway are divided into N mileage sections, and a grating sensing array is laid along the way; an edge server is deployed in each mileage section, and the data obtained by the grating sensing array within the mileage section is transmitted to the central server through a separate message queue MQ; N data structures are created in the central server to store the data transmitted by MQ respectively, and the data is aligned in time and space according to the spatio-temporal information obtained from the edge server. The same vehicle is searched for in the corresponding frame of each edge server in sequence, the vehicle information is merged, and the vehicle trajectories are stitched according to the mileage. The present invention realizes the identification of vehicle trajectories throughout the whole process and ensures the uniqueness of vehicles.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation, and in particular, to a method and system for stitching vehicle trajectories throughout a highway. Background Art

[0002] In recent years, with the development of sensing, communication, and information technologies, various countries have successively proposed development strategies for highway intelligentization, such as fifth-generation roads, cooperative intelligent transportation, connected vehicles, and smart highways, and have carried out a large number of technology R & D and demonstration promotion work. A smart highway is a dynamic system composed of the coordinated interaction of people, vehicles, roads, and the environment, and consists of several levels from bottom to top, including information perception, feature detection, pattern recognition, intelligent services, etc. Among them, vehicle feature detection and trajectory extraction are of great significance: on the one hand, it can support macroscopic highway traffic supervision and road planning functions, including traffic flow monitoring, traffic situation prediction, traffic event analysis, transportation planning, and road construction, etc.; on the other hand, it can provide microscopic highway intelligent services, including special vehicle tracking, vehicle behavior understanding, and complex environment analysis and vehicle trajectory prediction in the new generation of vehicle-road collaborative intelligent driving systems.

[0003] Currently, by deploying a grating sensing array fiber on each lane of the highway to construct a network information matrix, when a vehicle continuously moves on the road, it stimulates the sensing units along the line and generates a feedback signal sequence, and transmits the feedback signal sequence to the network information matrix to realize the detection of highway vehicle trajectories.

[0004] Due to the high sampling frequency of the grating sensing array fiber, the amount of data is huge, and the prior art cannot complete the recognition of vehicle trajectories throughout the highway; it cannot be flexibly expanded according to requirements. Summary of the Invention

[0005] Aiming at the defects existing in the prior art, the purpose of the present invention is to provide a method and system for stitching vehicle trajectories throughout a highway, realizing the recognition of vehicle trajectories throughout the highway and ensuring the uniqueness of vehicles.

[0006] To achieve the above object, on the one hand, a method for stitching vehicle trajectories throughout a highway is adopted, including:

[0007] Dividing the up and down directions of the highway into N mileage sections, and laying a grating sensing array along the way, where N > 1;

[0008] Deploying an edge server for each mileage section, and transmitting the data obtained by the grating sensing array within the mileage section to the central server through a separate message queue MQ;

[0009] Create N data structures in the central server to store the data transmitted by MQ respectively. Align the data in terms of time and space according to the spatio-temporal information obtained from the edge servers. Search for the same vehicle in each frame corresponding to the edge servers in sequence, merge the vehicle information, and splice the vehicle trajectories according to the mileage.

[0010] Preferably, configure the information of N edge servers in the configuration file of the central server, and then create N MQ clients, N data structures and corresponding N threads;

[0011] By increasing the number of edge server information, and then creating the same number of MQ clients, the same number of data structures and corresponding N threads, the expansion of the edge servers is realized.

[0012] Preferably, when the central server stores data, the data is automatically sorted according to the time stamp. The spatio-temporal alignment of the data includes the following steps in multiple rounds:

[0013] According to the time stamp of the first frame in the data structure, splice all the first data frames according to the mileage and lane direction, and then delete all the first data frames from the data structure.

[0014] Preferably, when the time stamps of all the first frames are within the error range, directly splice all the first data frames;

[0015] When the time stamp of the first frame in a data structure is outside the error range and the waiting times out, fill the corresponding first frame data with 0 and then splice all the first data frames.

[0016] Preferably, in the data center after spatio-temporal alignment, the data of each edge server is a spatio-temporally aligned frame; if in each round of splicing, there are vehicle trajectories that are not processed completely and there are the same vehicles in the same spatio-temporally aligned frame, then merge the same vehicles;

[0017] After merging or when there are no same vehicles in the same spatio-temporally aligned frame, if there are the same vehicles in other spatio-temporally aligned frames, then merge the information of the same vehicles and splice the vehicle trajectories according to the mileage;

[0018] After splicing or when there are no same vehicles in other spatio-temporally aligned frames, judge whether the vehicle trajectories in the next round are processed completely.

[0019] Preferably, the judgment basis for the same vehicle is: the time stamp is within the error range, in the same lane, the driving direction is the same, the mileage is within the error range and the vehicle type is the same.

[0020] The present invention also provides a system for splicing vehicle trajectories throughout the highway. The up and down directions of the highway are divided into N mileage sections. The system includes:

[0021] A grating sensing array is laid along the highway to obtain data feedback from vehicles;

[0022] N edge servers are respectively deployed at N mileage points to transmit the data within the corresponding mileage through separate message queues MQ;

[0023] A central server creates N data structures to respectively store the data transmitted by MQ, and is also used to spatially and temporally align the data according to the spatio-temporal information obtained from the edge servers, sequentially find the same vehicle in each frame corresponding to the edge server, merge the vehicle information, and splice the vehicle trajectories according to the mileage.

[0024] Preferably, N edge server information is configured in the central server configuration file to create N MQ clients, N data structures, and corresponding N threads;

[0025] By increasing the number of edge server information, and then creating the same number of MQ clients, the same number of data structures, and corresponding N threads, the expansion of the edge server is realized.

[0026] Preferably, when the central server stores data, the data is automatically sorted according to the time stamp. The spatio-temporal alignment of the data includes the following steps in multiple rounds:

[0027] According to the time stamp of the first frame in the data structure, after splicing all the first data frames according to the mileage and lane direction, all the first data frames are deleted from the data structure;

[0028] When the time stamps of all the first frames are within the error range, directly splice all the first data frames;

[0029] When the time stamp of the first frame in the data structure is outside the error range and the waiting times out, the corresponding first frame data is padded with 0 and then all the first data frames are spliced.

[0030] Preferably, in the data center after spatio-temporal alignment, the data of each edge server is a spatio-temporally aligned frame; if in each round of splicing, there is an unfinished vehicle trajectory and there are the same vehicles in the same spatio-temporally aligned frame, then the same vehicles are merged;

[0031] When there are no same vehicles after merging or in the same spatio-temporally aligned frame, if there are the same vehicles in other spatio-temporally aligned frames, then the same vehicle information is merged, and the vehicle trajectories are spliced according to the mileage;

[0032] After splicing or when there are no same vehicles in other spatio-temporally aligned frames, judge whether the next round of vehicle trajectory is processed completely.

[0033] One of the above technical solutions has the following beneficial effects:

[0034] Use edge servers to collect data from highway grating sensors, align the data of N edge servers spatiotemporally according to time, lane, and mileage, and then splice vehicle trajectories based on vehicle-road spatiotemporal information such as time, lane, mileage, vehicle speed, and vehicle length to ensure vehicle uniqueness; realize the identification of vehicle trajectories throughout the highway based on grating sensing technology.

[0035] In addition, configure the information of N edge servers in the central server configuration file. When expanding, only need to add the information of M edge servers in the central server configuration file to expand to N + M edge servers, and can be flexibly expanded according to requirements. Brief Description of the Drawings

[0036] Figure 1 It is a schematic diagram of vehicle trajectory splicing throughout the highway in the embodiment of the present invention;

[0037] Figure 2 It is a schematic diagram of the data structure of the central server in the embodiment of the present invention;

[0038] Figure 3 It is a flowchart of data spatiotemporal alignment in the embodiment of the present invention;

[0039] Figure 4 It is a flowchart for judging the same vehicle in the embodiment of the present invention. Detailed Embodiment

[0040] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0041] The present invention provides an embodiment of a method for splicing vehicle trajectories throughout a highway, including the following steps:

[0042] S1. Divide the up and down directions of the highway into N mileage sections, and lay a grating sensing array along the way, where N > 1.

[0043] S2. Deploy one edge server (EDGE) in each mileage section, and each edge server transmits the data obtained by the grating sensing array within that mileage section to the central server (CentralServer) through a separate MQ (Message Queue).

[0044] S3. Create N data structures in the central server to store the data transmitted by MQ respectively. Align the data in time and space according to the spatio-temporal information obtained from the edge servers. Search for the same vehicle in each frame corresponding to each edge server in turn, merge the vehicle information, and splice the vehicle trajectories according to the mileage.

[0045] For example Figure 1 is a schematic diagram of a specific architecture. The highway is divided into an up lane and a down lane. K0-K1, K1-K2, K2-K3, and K3-K4 are all one mileage, and each mileage corresponds to an edge server, and each edge server corresponds to a separate MQ. Since the up lane and the down lane are divided into the same number of mileages, Figure 1 there are a total of 8 mileages in, that is, N = 8, 8 edge servers and 8 MQs.

[0046] Further, in the above step S2, each edge server transmits the data obtained by the grating sensing array to the central server in parallel through the corresponding MQ at a fixed frequency, where the fixed frequency can be set according to the real-time requirement.

[0047] In the above step S3, configure the information of N edge servers in the central server configuration file, and then create N MQ clients, N data structures and corresponding N threads. In some embodiments, when expansion is needed, the number of edge server information can be added in the configuration file, and then the same number of MQ clients, the same number of data structures and the same number of N threads are created to achieve the expansion of the edge servers. For example, M edge server information can be added to the central server configuration file, and then M MQ clients, M data structures and corresponding M threads are created, and the system can be flexibly expanded to N + M edge servers, where M ≥ 1.

[0048] Further, create N MQ clients, N data structures and N corresponding threads in the central server according to the configuration file. The central server receives the data of N edge servers through N MQ clients respectively and stores them in the corresponding data structures. When storing the data, the data is automatically sorted according to the time stamp.

[0049] Combined with the attached Figure 1 , in some embodiments, within the above step S3, configure the JSON (JavaScript Object Notation) configuration file of 8 edge servers according to information such as mileage, IP, port, up and down directions, etc. The configuration file of the edge server is as follows:

[0050]

[0051]

[0052]

[0053] The data structure of the central server is as follows Figure 2 shown. The central server creates 8 MQ clients, 8 data structures and 8 corresponding threads according to the JSON configuration file. If 4 edge server configuration information is added to the JSON configuration file, and then 4 MQ clients, 4 data structures and 4 corresponding threads are created, it can be flexibly expanded to 12 edge servers.

[0054] As Figure 3 shown, in step S3 above, an embodiment of aligning data in time and space is provided, and the specific steps are as follows:

[0055] A301. The central server receives the data of N edge servers through MQ and stores them in the corresponding data structures respectively. The stored data is automatically sorted according to the time stamp.

[0056] A302. The central server retrieves the first frame from all data structures.

[0057] A303. Judge whether the time stamps of all the first frames are within the error range. If so, enter A304; if not, enter A306.

[0058] A304. Stitch all the first data frames according to mileage and lane direction.

[0059] A305. Delete all the first data frames from the data structure and transfer to A302.

[0060] A306. Judge whether the waiting times out. If so, enter A307; if not, transfer to A302.

[0061] A307. Pad the corresponding first frame data with 0 and transfer to A304.

[0062] In step S3 above, in the aligned data, the data of each edge server is a time-space alignment frame. If in each round of stitching, there is an unfinished vehicle trajectory and there are the same vehicles in the same time-space alignment frame, the same vehicles will be merged. After merging or when there are no same vehicles in the same time-space alignment frame, if there are the same vehicles in other time-space alignment frames, the same vehicle information will be merged and the vehicle trajectories will be stitched according to mileage; after stitching or when there are no same vehicles in other time-space alignment frames, judge whether the next round of vehicle trajectories is processed.

[0063] As Figure 4 shown, an embodiment of finding the same vehicle in stitching vehicle trajectories is provided, including the following process:

[0064] B301. Determine whether all vehicle trajectories in the splicing trajectory have been processed. If so, end; if not, go to B302.

[0065] B302. Determine whether there are the same vehicles in the same spatio-temporal alignment frame. If so, go to B303; if not, go to B304.

[0066] B303. Merge the information of the same vehicles.

[0067] B304. Determine whether there are the same vehicles in other spatio-temporal alignment frames. If so, go to B305; if not, transfer to B301.

[0068] B305. Merge the information of the same vehicles, splice the vehicle trajectories according to the mileage, and transfer to B301.

[0069] In the above process, it is determined whether it is the same vehicle according to the vehicle-road spatio-temporal information such as time stamp, lane, driving direction, mileage, speed, vehicle type, etc. In some embodiments, the judgment basis for the same vehicle is that the time stamp is within the error range, in the same lane, the driving direction is the same, the mileage is within the error range, and the vehicle type is the same. If any one of these conditions is not met, it does not belong to the same vehicle. After such judgment, the uniqueness of the estimated spliced vehicle can be guaranteed.

[0070] The present invention also provides an embodiment of a highway full-course vehicle trajectory splicing system, which can implement the above method. As Figure 1 shown, the system includes a grating sensing array, N edge servers, and a central server. The grating sensing array is laid along the highway for obtaining the data fed back by vehicles, where the up and down directions of the highway are divided into N mileage sections, and the length of each mileage section is equal. The N edge servers are respectively deployed in the N mileage sections for transmitting the data within the corresponding mileage sections through separate message queues MQ, that is, each mileage section corresponds to one edge server and one MQ.

[0071] The central server creates N data structures for respectively storing the data transmitted by MQ, and is also used for spatio-temporally aligning the data according to the spatio-temporal information obtained from the edge servers, sequentially searching for the same vehicles in the frames corresponding to each edge server, merging the vehicle information, and splicing the vehicle trajectories according to the mileage.

[0072] The central server is also used to configure the information of the N edge servers in the configuration file, and then create N MQ clients, N data structures, and corresponding N threads. By increasing the number of edge server information, and then creating the same number of MQ clients, the same number of data structures, and corresponding N threads, the expansion of the edge servers is realized.

[0073] When the central server stores data, it automatically sorts the data according to the timestamp, and the steps for spatio-temporal alignment of the data include the following steps for multiple rounds:

[0074] According to the timestamp of the first frame in the data structure, after splicing all the first data frames according to the mileage and lane direction, all the first data frames are deleted from the data structure.

[0075] Among them, when the timestamps of all the first frames are within the error range, all the first data frames are directly spliced; when the timestamp of the first frame in the data structure is outside the error range and the waiting times out, the corresponding first frame data is padded with 0 and then all the first data frames are spliced.

[0076] In the data center after the above spatio-temporal alignment, the data of each edge server is a spatio-temporal alignment frame; if in each round of splicing, there is an unfinished vehicle trajectory and there are the same vehicles in the same spatio-temporal alignment frame, the same vehicles are merged; after the merger or when there are no same vehicles in the same spatio-temporal alignment frame, if there are the same vehicles in other spatio-temporal alignment frames, the same vehicle information is merged, and the vehicle trajectories are spliced according to the mileage; after the splicing or when there are no same vehicles in other spatio-temporal alignment frames, it is judged whether the next round of vehicle trajectories is processed.

[0077] The present invention deploys N edge servers, then aligns and splices the data according to the spatio-temporal characteristics of the vehicle-road, and ensures the uniqueness of the vehicle through the trajectory recognition algorithm according to the spatio-temporal characteristics of the vehicle-road, so as to realize the splicing of the vehicle trajectories throughout the expressway. The flexible expansion of the edge servers is realized by configuring the information of N edge servers in the configuration file of the central server.

[0078] The above are only the embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention are included in the scope of the claims of the present invention pending approval.

Claims

1. A method for stitching vehicle trajectories throughout a highway, characterized in that, Including: Dividing the up and down directions of the highway into N mileage sections, and laying a grating sensing array along the way, where N > 1; Deploying one edge server for each mileage section, and transmitting the data obtained by the grating sensing array within that mileage section to the central server through a separate message queue MQ; Creating N data structures in the central server to store the data transmitted by MQ respectively, aligning the data in space and time according to the spatio-temporal information obtained from the edge servers, sequentially searching for the same vehicle in the frames corresponding to each edge server, merging the vehicle information, and splicing the vehicle trajectories according to the mileage; When the central server stores data, it automatically sorts the data according to the time stamp. The steps for aligning the data in space and time include multiple rounds as follows: According to the time stamp of the first frame in the data structure, after splicing all the first data frames according to the mileage and lane direction, delete all the first data frames from the data structure; When the time stamps of all the first frames are within the error range, directly splice all the first data frames; When the time stamp of the first frame in a data structure is outside the error range and the waiting times out, fill the corresponding first frame data with 0 and then splice all the first data frames.

2. The method for stitching vehicle trajectories throughout a highway according to claim 1, wherein Configuring the information of N edge servers in the configuration file of the central server, and then creating N MQ clients, N data structures and corresponding N threads; By increasing the number of edge server information, and then creating the same number of MQ clients, the same number of data structures and corresponding N threads, the expansion of the edge server is realized.

3. The method for splicing vehicle trajectories throughout a highway according to claim 1, wherein, In the data center after the spatio-temporal alignment, the data of each edge server is a spatio-temporally aligned frame; if in each round of splicing, there are vehicle trajectories that are not processed completely, and there are the same vehicles in the same spatio-temporally aligned frame, then merge the same vehicles; After merging or when there are no same vehicles in the same spatio-temporally aligned frame, if there are the same vehicles in other spatio-temporally aligned frames, then merge the information of the same vehicles and splice the vehicle trajectories according to the mileage; After splicing or when there are no same vehicles in other spatio-temporally aligned frames, judge whether the vehicle trajectories in the next round are processed completely.

4. The method for stitching vehicle trajectories throughout a highway according to claim 3, wherein, The basis for judging the same vehicle is: the time stamp is within the error range, in the same lane, the driving direction is the same, the mileage is within the error range and the vehicle type is the same.

5. A vehicle trajectory stitching system for the entire length of a highway, characterized in that, The up and down directions of the highway are divided into N mileage sections, and the system includes: A grating sensing array, laid along the highway, for obtaining the data fed back by vehicles; N edge servers, respectively deployed in N mileage sections, for transmitting the data within the corresponding mileage section through a separate message queue MQ; A central server, creating N data structures, for storing the data transmitted by MQ respectively, and also for aligning the data in space and time according to the spatio-temporal information obtained from the edge servers, sequentially searching for the same vehicle in the frames corresponding to each edge server, merging the vehicle information, and splicing the vehicle trajectories according to the mileage; When the central server stores data, it automatically sorts the data according to the time stamp. The steps for aligning the data in space and time include multiple rounds as follows: According to the time stamp of the first frame in the data structure, after splicing all the first data frames according to the mileage and lane direction, delete all the first data frames from the data structure; When the timestamps of all the first frames are within the error range, directly splice all the first data frames; When the timestamp of the first frame in the data structure is outside the error range and the waiting times out, pad the corresponding first frame data with 0 and then splice all the first data frames.

6. The vehicle trajectory stitching system for the entire highway as described in claim 5, characterized in that, Configure the information of N edge servers in the central server configuration file, and create N MQ clients, N data structures and the corresponding N threads; Realize the expansion of edge servers by increasing the number of edge server information, and then creating the same number of MQ clients, the same number of data structures and the corresponding N threads.

7. The vehicle trajectory stitching system for the entire highway as described in claim 5, characterized in that, In the data center after spatio-temporal alignment, the data of each edge server is a spatio-temporal alignment frame; if there are vehicle trajectories that are not processed completely in each round of splicing and there are the same vehicles in the same spatio-temporal alignment frame, merge the same vehicles; When there are no same vehicles after merging or in the same spatio-temporal alignment frame, if there are the same vehicles in other spatio-temporal alignment frames, merge the information of the same vehicles and splice the vehicle trajectories according to the mileage; After splicing or when there are no same vehicles in other spatio-temporal alignment frames, judge whether the vehicle trajectories in the next round are processed completely.

Images