CN117473741A - Full-sample high-resolution vehicle track robust reconstruction method, device and medium - Google Patents
Full-sample high-resolution vehicle track robust reconstruction method, device and medium Download PDFInfo
- Publication number
- CN117473741A CN117473741A CN202311425411.0A CN202311425411A CN117473741A CN 117473741 A CN117473741 A CN 117473741A CN 202311425411 A CN202311425411 A CN 202311425411A CN 117473741 A CN117473741 A CN 117473741A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- track
- time
- networked
- candidate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 88
- 239000011159 matrix material Substances 0.000 claims abstract description 26
- 238000011144 upstream manufacturing Methods 0.000 claims abstract description 18
- 230000000903 blocking effect Effects 0.000 claims abstract description 10
- 238000009499 grossing Methods 0.000 claims abstract description 10
- 230000006872 improvement Effects 0.000 claims abstract description 6
- 230000008569 process Effects 0.000 claims description 20
- 230000001133 acceleration Effects 0.000 claims description 12
- 238000001514 detection method Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000035484 reaction time Effects 0.000 claims description 3
- 230000002123 temporal effect Effects 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 238000007499 fusion processing Methods 0.000 abstract 1
- 230000035699 permeability Effects 0.000 description 13
- 230000000694 effects Effects 0.000 description 8
- 230000003993 interaction Effects 0.000 description 8
- 230000006399 behavior Effects 0.000 description 7
- 230000004927 fusion Effects 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 241000764238 Isis Species 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000009514 concussion Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000452 restraining effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000000087 stabilizing effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/04—Constraint-based CAD
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to a full-sample high-resolution vehicle track robust reconstruction method, device and medium, wherein the reconstruction method comprises the following steps: acquiring coil detector data and network connection data, and estimating a space-time velocity matrix by a self-adaptive smoothing method based on traffic jam flow and free flow characteristic improvement; selecting a non-network vehicle to be reconstructed, and generating candidate tracks through an IDM model according to the upstream and downstream known tracks; calculating the weight of the candidate track by taking the generated space-time velocity matrix as constraint on the non-network vehicle connected with the generated candidate track, fusing the candidate track by using a weighting method, and reconstructing the high-resolution vehicle track of the non-network vehicle connected with the vehicle; and sequentially repeating the candidate track generation and fusion process for each non-networked vehicle. Compared with the prior art, the invention improves the reconstruction precision and robustness; the track information of the vehicle in different scenes of the blocking flow and the free flow is accurately estimated, and the optimal reconstruction result can be approached through simple parameter adjustment.
Description
Technical Field
The invention relates to the technical field of intelligent transportation, in particular to a full-sample high-resolution vehicle track robust reconstruction method, device and medium with scene flexibility.
Background
The high-resolution vehicle track can provide a large amount of traffic space-time information, can extract vehicle group characteristics such as flow, speed, density and the like, can be used for analyzing microscopic behaviors of interaction among vehicles, and provides a good basis for traffic state estimation, traffic flow modeling, signal control optimization, traffic emission measurement and the like. However, the conventional fixed detector is not high in deployment rate, and the mobile detector is low in permeability and cannot be used for directly acquiring high-resolution vehicle tracks; the cost for directly acquiring the high-resolution track through unmanned aerial vehicle shooting and other means is high. In the context of difficult trajectory acquisition, many vehicle trajectory reconstruction methods have developed, but the following problems remain: the model assumption is inconsistent with reality, the data requirement is high, the design scene is single, and the practical application of the methods is hindered. Therefore, the vehicle track reconstruction model which is reasonable in assumption, suitable for various data conditions and flexibly corresponding to traffic scenes is constructed by fully fusing and utilizing the data detected by various sensors, and the complete high-resolution vehicle track data is reconstructed from sparse data observation, so that the vehicle track reconstruction model has great significance to the intelligent traffic field.
The early vehicle track reconstruction method is limited by the detector technology, mainly utilizes fixed detector data (coils, radars, license plate recognition and the like) to reconstruct tracks through methods such as variation theory, traffic flow basic diagrams and the like, and the reconstruction result is mainly used for vehicle travel time estimation and the like. The method ignores complex behaviors such as acceleration, deceleration, following and the like of the vehicle to a great extent, does not have the capability of describing microscopic behaviors among vehicles, and cannot be used for microscopic applications (traffic safety, traffic concussion research and the like). With the development of motion detectors, some methods use vehicle-following models to reconstruct trajectories based on networked and autonomous vehicle data, often assuming that networked and autonomous vehicles have higher permeabilities. But the permeability of the internet-connected and automatic driving vehicles in the actual scene is basically below 10%, and the permeability is difficult to be greatly improved in a short time.
The prior art has the following defects:
(1) Most of the existing methods can not solve the problem that the microscopic behavior of the vehicle is characterized and the sparse data scene is compatible. The method based on the fixed point detection data has low data requirement, but can not accurately show vehicle behaviors such as acceleration and deceleration; the method based on the data of the mobile detector better characterizes microscopic behaviors of the vehicle, but has high requirements on data quality, and is supposed to contradict the actual scene on the premise of high uploading frequency and high permeability.
(2) In the existing method, multi-source data fusion is insufficient. Although a framework for restraining the reconstruction of the microscopic track in the macroscopic traffic state is proposed, interaction among the macroscopic and microscopic modules is insufficient, the data fusion degree and the data use efficiency are not high, and the accuracy of the reconstruction of the vehicle track needs to be further improved.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a full-sample high-resolution vehicle track robust reconstruction method with scene flexibility, wherein an error impedance model (ER) is introduced in the aspect of data fusion, so that interaction between macroscopic speed information and microscopic track information is increased, multi-source data is more fully utilized, and reconstruction precision and robustness are greatly improved; on the other hand, the track information of the vehicle in different scenes of the choked flow and the free flow is accurately estimated by using an Intelligent Driver Model (IDM), and the optimal reconstruction result can be accessed through simple parameter adjustment.
The aim of the invention can be achieved by the following technical scheme:
the first aspect of the invention provides a full-sample high-resolution vehicle track robust reconstruction method, which comprises the following steps:
s1: acquiring coil detector data and network connection data, and estimating a space-time velocity matrix by a self-adaptive smoothing method based on traffic jam flow and free flow characteristic improvement;
s2: selecting a non-network vehicle with a track to be reconstructed, determining an upstream reference track and a downstream reference track of the vehicle based on an error impedance model (ER), and generating a candidate track of the vehicle through an IDM model according to the reference track;
s3: and (2) for the non-network connected vehicle with the candidate track generated in the step (S2), calculating weights of the two candidate tracks by taking the space-time velocity matrix generated in the step (S1) as constraint, fusing the two candidate tracks by a weighting method to reconstruct the track of the non-network connected vehicle, and returning to the step (S2) until all the non-network connected vehicle tracks are sequentially reconstructed.
High resolution vehicle trajectory refers to data describing vehicle position and motion in fine granularity in time and space.
Further, in S1, the coil detector data includes: detecting record ID, vehicle instantaneous speed and vehicle elapsed time;
the internet protocol data comprises: vehicle ID, timestamp, vehicle coordinates, vehicle instantaneous speed are detected.
Further, the acquiring process of the space-time velocity matrix comprises the following steps:
s1-1: for a time-space domain on a single lane, dividing a space-time velocity matrix by taking a 3-meter space and a 3-second time as minimum set counting units, and initializing the space-time velocity matrix by using the known traffic velocity v at the time of an x position t in coil detector data and network vehicle connection data;
s1-2: calculating a smoothing kernel phi (,) and a normalization factor for each matrix element (x, t) of unknown numerical value
S1-3: simultaneously, the traffic blocking flow and the free flow characteristics are considered to respectively adjust the smooth kernel V free (x, t) and V cong (x，t)；
S1.4: calculating weights w (x, t) to trade-off free flow and blocked flow characteristics;
s1-5: estimating an unknown traffic speed under space-time (x, t), and complementing the space-time speed matrix:
V refer (x，t)＝w(x，t)V cong (x，t)+[1-w(x，t)]V free (x，t)。
further, in S1-2, the smoothing kernel φ (-) and the normalization factorThe acquisition mode of (a) is as follows:
wherein x is i 、t i 、v i (i=1,..n.) is the known position, time and traffic speed in the corresponding space-time, respectively, the smoothed width σ in the spatial coordinates is 80m and the smoothed width τ in the temporal coordinates is 6.5s;
in S1-3, smooth kernel V is adjusted free (x, t) and V cong (x, t) is:
wherein the propagation speed c of the traffic disturbance in the free flow free Taking the propagation speed c of traffic disturbance in the blocking flow at 70km/h cong Taking-15 km/h;
in S1-4, the calculation process of the weight w (x, t) is as follows:
wherein the threshold value V between free flow and blocked flow thr The transition width DeltaV between the free stream and the choked stream was taken at 60km/h and 20km/h.
Further, in S2, the specific process is as follows:
s2-1: i-piece internet protocol (NET) is shared in internet protocol vehicle track set YThe vehicle track obtains the adjacent upstream network vehicle-connected track Y i And downstream net train track Y i+1 Reconstruction interval therebetweenWherein N represents the number of non-networked vehicle tracks to be reconstructed in the interval, and i=1 in the initial state;
s2-2: for the intervalThe nth non-network vehicle in the system determines the upstream reference track X of the vehicle upREF And a downstream reference trajectory X downREF N=1 in the initial state;
s2-3: based on IDM model, according to upstream reference track X upREF Generating candidate tracks of the nth non-networked vehicle
S2-4: based on IDM model, according to downstream reference track X downREF Generating candidate tracks of the nth non-networked vehicle
Further, in S2-2, the upstream reference trajectory X upREF And a downstream reference trajectory X downREF The determining process of (1) comprises:
a. if n=1, x upREF ＝Y i ，
b. If n=n,X downREF ＝Y i+1 ；
c. if N is not equal to 1 and N is not equal to N,
wherein the method comprises the steps ofIs the reconstruction track of the n-1 non-network vehicle; />Is the candidate track +.>Candidate trajectories generated for reference, +.>Is a downstream network-connected track Y i+1 Candidate trajectories generated for the reference;
s2-3 according to the upstream reference track X upREF Generating candidate tracks of the nth non-networked vehicleThe process of (1) is as follows:
a. from the following componentsCalculate->
b. From the following componentsCalculate->
min location error＝|x′ upREF (t-2)-x upREF (t-2)|
Wherein, initiallyIs provided by coil detection data->And->Acceleration, speed and position at time t estimated by non-networked vehicle n with reference to preceding vehicle, v upREF (t) and x upREF (t) the speed and position of the preceding vehicle of the non-networked vehicle n at the moment t, a ', v ' and x ' are estimated values, and the maximum acceleration a takes 2.75m/s 2 The most comfortable deceleration b takes 2.25m/s 2 Free flow vehicle speed v 0 Taking 32m/s, s 0 The distance between safety vehicles is 8m, s * The vehicle distance is 1.1s for the reaction time T.
Further, in S2-4, the reference trajectory X is determined based on the downstream downREF Generating candidate tracks of the nth non-networked vehicleThe process of (1) is as follows:
a. from the following componentsCalculate->
minlocation error＝|x′ downREF (t+2)-x downREF (t+2)|
b. From the following componentsCalculate->
Wherein, initiallyIs provided by coil detection data->And->Acceleration, speed and position at time t, v, estimated by non-networked vehicle n with reference to rear vehicle downREF (t) and x downREF (t) speed and position of the following vehicle of the non-networked vehicle n at the time t.
Further, in S3, the specific process is:
s3-1: the space-time velocity matrix V estimated in S1 refer As constraint, solving candidate track of non-internet-connected vehicle n in S2And->Weight of +.>And->
S3-2: high-resolution track of non-internet-connected vehicle n is calculated according to weighting method
S3-3: returning to S2 to reconstruct the track of the next non-networked vehicle until the reconstruction intervalThe N non-networked vehicles in (a) are all reconstructed, i.e., n=n.
S3-4: let i=i+1, n=1, return S2 to continue reconstructing the trajectories of N non-networked vehicles in the next section until all non-networked vehicles in the section are reconstructed, i.e. i=i-1.
A second aspect of the present invention provides an electronic device, including a memory, and a processor, where the processor is configured to execute a program in the memory, so as to implement a full-sample high-resolution vehicle track robust reconstruction method as described above.
A third aspect of the present invention provides a storage medium containing computer executable instructions which, when executed by a computer processor, are used to perform a full-sample high-resolution vehicle trajectory robust reconstruction method as described above.
Compared with the prior art, the invention has the following technical advantages:
(1) In the existing method, although the reconstruction of the track is constrained by macroscopic traffic information, the utilization of the macroscopic information is insufficient, so that errors are accumulated along with the increase of the number of the non-network vehicles to be reconstructed, and the method lacks robustness. The proposed method enables interaction of macroscopic and microscopic traffic information to be more complete, and robustness of high-resolution track reconstruction is increased.
(2) The existing method also combines traffic macroscopic and microscopic models, but the Newell model of the microscopic module in the existing method assumes that the tracks of the front and rear vehicles are consistent, and lacks accurate description of interaction between vehicles, so that the reconstructed track is distorted under the blocking flow with low permeability. The method is compatible with vehicle track reconstruction in a blocking flow scene and a free flow scene, and the description of microscopic vehicle behaviors is more reasonable and accurate.
Drawings
FIG. 1 is a graph showing the effect of an IDM model before and after use at 5% on-line permeability;
FIG. 2 is a graph showing the effect of the ER model under 5% Internet protocol permeation before and after use;
FIG. 3 shows the errors of the proposed method under different parameter settings;
fig. 4 is an overall flow chart in the present embodiment.
Detailed Description
The invention provides a full-sample high-resolution track robust reconstruction method with scene flexibility, which comprises the following steps: under the macro-micro module interaction framework of an error impedance model (ER), firstly estimating a macro space-time velocity matrix by using sparse network connection data and easily acquired coil detector data, and then generating a micro vehicle track by using an Intelligent Driver Model (IDM) with macro velocity information as constraint. On one hand, the ER model is introduced in the aspect of data fusion, so that interaction between macroscopic speed information and microscopic track information is increased, and reconstruction accuracy and robustness are greatly improved; on the other hand, the IDM model is used for accurately estimating the track information of the vehicle in different scenes of the blocking flow and the free flow, and the optimal reconstruction result can be approached through simple parameter adjustment.
The invention will now be described in detail with reference to the drawings and specific examples. Features such as a part model, a material name, a connection structure, a control method, an algorithm and the like which are not explicitly described in the technical scheme are all regarded as common technical features disclosed in the prior art.
Example 1
The full-sample high-resolution vehicle track robust reconstruction method with scene flexibility in the embodiment comprises the following steps, see fig. 4:
step 1: coil detector data and network connection data are input, and a space-time velocity matrix is estimated through an Adaptive Smoothing Method (ASM) based on traffic jam flow and free flow characteristic improvement, namely a traffic macroscopic model.
Step 2: and selecting a non-network vehicle with a track to be reconstructed, determining an upstream reference track and a downstream reference track of the vehicle based on an error impedance model (ER), and generating a candidate track of the vehicle through an IDM model (traffic micro model) according to the reference track.
Step 3: calculating weights of two candidate tracks by taking the space-time velocity matrix generated in the step (1) as constraint on the non-network vehicle with the generated candidate tracks in the step (2), and fusing the two candidate tracks by a weighting method to reconstruct the track of the non-network vehicle; and returning to the step 2 until all the non-internet-connected vehicle tracks are sequentially reconstructed.
Wherein IDM (Intelligent Driver Model) is an intelligent driver model that has a small number of parameters, is well defined, and can describe different states from free flow to fully congested flow with a unified model.
The specific process of the step 1 is as follows:
(1) Inputting coil detector data including a detection record ID, vehicle instantaneous speed, vehicle elapsed time; network connection data including detected vehicle ID, time stamp, vehicle coordinates, vehicle instantaneous speed is input.
(2) And dividing a space-time velocity matrix by taking 3 m space and 3 s time as minimum set counting units for a time-space domain on a single lane, and initializing the matrix by using the recorded traffic velocity v at the time of the x position t in the coil and network vehicle connection data.
(3) Calculating a smoothing kernel phi (,) and a normalization factor for each matrix element (x, t) of unknown numerical value
Wherein x is i 、t i 、v i (i=1,..n.) is the known position, time and traffic speed in the corresponding space-time, respectively, the smoothed width σ in the spatial coordinates is 80m and the smoothed width τ in the temporal coordinates is 6.5s.
(4) Simultaneously considering traffic blocking flow and free flow characteristics, the smoothing kernels are respectively adjusted to be:
wherein the propagation speed c of the traffic disturbance in the free flow free Taking the propagation speed c of traffic disturbance in the blocking flow at 70km/h cong Taking-15 km/h.
(5) Calculating weights w (x, t), weigh the free stream and the blocked stream features:
wherein the threshold value V between free flow and blocked flow thr The transition width DeltaV between the free stream and the choked stream was taken at 60km/h and 20km/h.
(6) Estimating an unknown traffic speed under space-time (x, t), and complementing the space-time speed matrix:
V refer (x，t)＝w(x，t)V cong (x，t)+[1-w(x，t)]V free (x，t)
the specific process of the step 2 is as follows:
(1) The network-connected vehicle track set Y shares I network-connected vehicle tracks, and adjacent upstream network-connected vehicle tracks Y i And downstream net train track Y i+1 With a reconstruction interval therebetweenN represents the number of non-networked vehicle tracks that need to be reconstructed in this interval, i=1 in the initial condition.
(2) For the intervalThe nth non-network vehicle in the system determines the upstream reference track X of the vehicle upREF And a downstream reference trajectory X downREF N=1 under initial conditions:
a. if n=1, x upREF ＝Y i ，
b. If n=n,X downREF ＝Y i+1 ；
c. if N is not equal to 1 and N is not equal to N,
wherein the method comprises the steps ofThe reconstruction track of the n-1 non-network vehicle can further reduce error generation by utilizing information in the reconstructed track, which is called an error impedance model (ER); />Is the candidate track +.>(n+2 < N) generating candidate trajectories for reference,>is a downstream network-connected track Y i+1 Candidate trajectories are generated for the reference.
(3) Based on IDM model, according to upstream reference track X upREF Generating candidate tracks of the nth non-networked vehicle
a. From the following componentsCalculate->
b. From the following componentsCalculate->
min location error＝|x′ upREF (t-2)-x upREF (t-2)|
Wherein initially there isIs provided by coil detection data->And->Acceleration, speed and position at time t estimated by non-networked vehicle n with reference to preceding vehicle, v upREF (t) and x upREF (t) speed and position of the lead vehicle (i.e., the upstream reference vehicle) of the non-networked vehicle n at time t, a ', v ', and x ' are estimates. Maximum acceleration a of 2.75m/s 2 The most comfortable deceleration b takes 2.25m/s 2 Free flow vehicle speed v 0 Taking 32m/s, s 0 The distance between safety vehicles is 8m, s * The vehicle distance is 1.1s for the reaction time T.
(4) Based on IDM model, according to downstream reference track X downREF Generating candidate tracks of the nth non-networked vehicle
a. From the following componentsCalculate->
min location error＝|x′ downREF (t+2)-x downREF (t+2)|
b. From the following componentsCalculate->
Wherein initially there isIs provided by coil detection data->And->Acceleration, speed and position at time t, v, estimated by non-networked vehicle n with reference to rear vehicle downREF (t) and x downREF (t) speed and position of a following vehicle (i.e., a downstream reference vehicle) that is not a networked vehicle n at time t.
The specific process of the step 3 is as follows:
(1) The space-time velocity matrix V estimated in the step 1 refet As a constraint, solving the candidate track of the non-internet-connected vehicle n in the step 2And->Weight of +.>And->
(2) High-resolution track of non-internet-connected vehicle n is calculated according to weighting method
(3) Returning to S2 to continuously calculate the candidate track of the next non-networked vehicle until the reconstruction intervalThe N non-networked vehicles in (a) are all reconstructed, i.e., n=n.
(4) Let i=i+1, n=1, return to S2 to continue to calculate candidate trajectories for N non-networked vehicles in the next reconstruction interval until non-networked vehicles in all intervals are reconstructed, i.e. i=i-1.
Verification example 1
The method performs case analysis on the method, and has good track reconstruction effect. The case data is from the NGSIM dataset collected by the federal public administration in the united states. Vehicle track data captured by unmanned aerial vehicle shooting between 7 minutes and 7 minutes at 7.m. in the leftmost lane running from north to south in the expressway 101 los angeles of california in the united states is analyzed.
In order to verify the compatibility of the method to a low-permeability scene, a sparse data scene with 85%, 90% and 95% of network-connected vehicle data missing is constructed to test performance, and a Root Mean Square Error (RMSE), a Mean Absolute Error (MAE) and a Mean Absolute Percentage Error (MAPE) are used as indexes for measuring the data recovery precision, wherein the calculation method is as follows:
wherein the method comprises the steps ofIs the observation trace point, +.>Is the reconstructed result trace point. MAE represents the average position error, RMSE is sensitive to outliers and extrema, and MAPE is the size of the error relative to the length of the reconstructed road segment.
The performance of the method combining the macro-micro model is superior to that of the traditional interpolation and variation method, and macro-micro modules based on different models are selected to test the performance in order to illustrate the superiority of the proposed method. The test results are shown in Table 1. As the permeability increases from 5% to 15%, and particularly from 5% to 10%, the error in the representation of the three indicators decreases. It is stated that permeability significantly affects the performance of the various reconstruction methods, and that high resolution vehicle trajectory reconstruction at very low permeability is challenging. Under these low permeability scenes, the proposed method not only reconstructs more accurate vehicle trajectories (reconstruction accuracy at each permeability is improved by more than 10%), but also has better robustness to sparse data scenes (reconstruction accuracy at 5% permeability is improved by 28.9%).
Table 1 vehicle track drop member energy for different macro-micro fusion methods
The introduction of the IDM model allows the method to have scene flexibility. Newell is a widely used following model and figure 1 compares the reconstruction effect using Newell and IDM models in the microscopic model. The method used by the invention more accurately reconstructs the process that the vehicle is decelerated to slow running when the vehicle flow is blocked, and is accelerated to recover to the free flow speed after a period of time, and simultaneously ensures the reconstruction effect of the vehicle track in the free flow state.
An error impedance model (ER) is introduced in the process of macro-micro information interaction to reduce errors generated in the track reconstruction process. Fig. 2 compares the errors of the candidate trajectories generated in step S2 before and after using the ER model. As shown in fig. 2 (a), the error increases with the increase of the number of non-network-connected vehicles that need to be reconstructed between adjacent network-connected vehicle tracks, that is, the larger the non-network-connected vehicle reconstruction error is, the farther the non-network-connected vehicle is from the network-connected vehicle, and the ER model reduces the error; fig. 2 (b) and 2 (c) intuitively reflect the effect of the ER model, and the candidate trajectories generated by the model are closer to the true value, both in the free-flow and choked flow states. These improvements are due to the fact that when candidate trajectory estimation is performed based on the ER model, the vehicle trajectory that the preceding vehicle has reconstructed is used as a reference, and the reconstructed trajectory of the preceding vehicle is generated under the constraint of macroscopic traffic speed information. In this way, traffic macroscopic speed information is more fully utilized, and reconstruction errors are further reduced.
To measure the effect of parameters on model performance, the invented method was subjected to parameter sensitivity test, and the results are shown in fig. 3. When the free flow velocity v 0 Setting the average absolute error to be 20-36m/s, and stabilizing the average absolute error to be about 8 m; when the safety distance is set to 6-9m, the average absolute error is stabilized below 9 m. The performance of the model is also more stable around the optimal parameter values for maximum acceleration a and comfort deceleration b. The method is insensitive to parameters, can approach to the optimal effect through simple parameter adjustment, and has competitiveness in practical application.
The embodiment also provides full-sample high-resolution vehicle track robust reconstruction equipment, which comprises a processor and a memory, wherein the processor is coupled with the memory, the memory stores program instructions, and the task management method is realized when the program instructions stored in the memory are executed by the processor. The processor may be a general-purpose processor, including a Central Processing Unit (CPU), a network processor (Network Processor NP), and the like; but also Digital Signal Processors (DSP), application Specific Integrated Circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components; the memory may comprise Random Access Memory (RAM) or Non-volatile memory (Non-volatile memory), such as at least one disk memory. The memory may be an internal memory of the random access memory (RandomAccessMemory, RAM) type, and the processor, memory may be integrated as one or more separate circuits or hardware, such as: an application specific integrated circuit (ApplicationSpecificIntegratedCircuit, ASIC). It should be noted that the computer program in the above-mentioned memory may be implemented in the form of software functional units and may be stored in a computer readable storage medium when sold or used as a separate product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, an electronic device, or a network device, etc.) to perform all or part of the steps of the method of the various embodiments of the present invention.
The present embodiment also proposes a computer-readable storage medium storing computer instructions for causing a computer to execute the above-described full-sample high-resolution vehicle trajectory robust reconstruction method. The storage medium may be an electronic medium, a magnetic medium, an optical medium, an electromagnetic medium, an infrared medium, or a semiconductor system or propagation medium. The storage medium may also include semiconductor or solid state memory, magnetic tape, removable computer diskette, random Access Memory (RAM), read-only memory (ROM), rigid magnetic disk and optical disk. Optical discs may include compact disc-read only memory (CD-ROM), compact disc-read/write (CD-RW), and DVD.
The previous description of the embodiments is provided to facilitate a person of ordinary skill in the art in order to make and use the present invention. It will be apparent to those skilled in the art that various modifications can be readily made to these embodiments and the generic principles described herein may be applied to other embodiments without the use of the inventive faculty. Therefore, the present invention is not limited to the above-described embodiments, and those skilled in the art, based on the present disclosure, should make improvements and modifications without departing from the scope of the present invention.
Claims (10)
1. The full-sample high-resolution vehicle track robust reconstruction method is characterized by comprising the following steps of:
s1: acquiring coil detector data and network connection data, and estimating a space-time velocity matrix by a self-adaptive smoothing method based on traffic jam flow and free flow characteristic improvement;
s2: selecting a non-network vehicle with a track to be reconstructed, determining an upstream reference track and a downstream reference track of the vehicle based on an error impedance model (ER), and generating a candidate track of the vehicle through an IDM model according to the reference track;
s3: and (3) for the non-network connected vehicle with the candidate track generated in the step (S2), calculating weights of the two candidate tracks by taking the space-time velocity matrix generated in the step (S1) as constraint, fusing the two candidate tracks by using a weighting method to reconstruct the track of the non-network connected vehicle, and returning to the step (S2) until all the non-network connected vehicle tracks are sequentially reconstructed.
2. The full-sample high-resolution vehicle trajectory robust reconstruction method of claim 1, wherein in S1, said coil detector data comprises: detecting record ID, vehicle instantaneous speed and vehicle elapsed time;
the internet protocol data comprises: vehicle ID, timestamp, vehicle coordinates, vehicle instantaneous speed are detected.
3. The full-sample high-resolution vehicle trajectory robust reconstruction method of claim 1, wherein the space-time velocity matrix acquisition process comprises:
s1-1: dividing a space-time velocity matrix by taking 3 m space and 3 s time as minimum set units on a time-space domain on a single lane, and initializing the space-time velocity matrix by using recorded traffic velocity v at the time of an x position t in coil detector data and network vehicle connection data;
s1-2: calculating a smoothing kernel phi (,) and a normalization factor for each matrix element (x, t) of unknown numerical value
S1-3: simultaneously, the traffic blocking flow and the free flow characteristics are considered to respectively adjust the smooth kernel V free (x, t) and V cong (x，t)；
S1-4: calculating weights w (x, t) to trade-off free flow and blocked flow characteristics;
s1-5: estimating an unknown traffic speed under space-time (x, t), and complementing the space-time speed matrix:
V refer (x，t)＝w(x，t)V cong (x，t)+[1-w(x，t)]V free (x，t)。
4. a full-sample high-resolution vehicle trajectory robust reconstruction method as recited in claim 3, wherein in S1-2, said smoothing kernel Φ (·) and normalization factorThe acquisition mode of (a) is as follows:
wherein x is i 、t i 、v i (i=1,..n.) is the known position, time and traffic speed in the corresponding space-time, respectively, the smoothed width σ in the spatial coordinates is 80m and the smoothed width τ in the temporal coordinates is 6.5s;
in S1-3, smooth kernel V is adjusted free (x, t) and V cong (x, t) is:
wherein the propagation speed c of the traffic disturbance in the free flow free Taking the propagation speed c of traffic disturbance in the blocking flow at 70km/h cong Taking-15 km/h;
in S1-4, the calculation process of the weight w (x, t) is as follows:
wherein the threshold value V between free flow and blocked flow thr The transition width DeltaV between the free stream and the choked stream was taken at 60km/h and 20km/h.
5. A full-sample high-resolution vehicle track robust reconstruction method as recited in claim 3, wherein in S2, the specific process is as follows:
s2-1: i net train tracks are shared in the net train track set Y, and adjacent upstream net train tracks Y are obtained i And downstream net train track Y i+1 Reconstruction interval therebetweenWherein N represents the number of non-networked vehicle tracks to be reconstructed in the interval, and i=1 under the initial condition;
s2-2: for the intervalThe nth non-network vehicle in the system determines the upstream reference track X of the vehicle upREF And a downstream reference trajectory X downREF N=1 under initial conditions;
s2-3: based on IDM model, according to upstream reference track X upREF Generating candidate tracks of the nth non-networked vehicle
S2-4: based on IDM model, according to downstream reference track X downREF Generating candidate tracks of the nth non-networked vehicle
6. The full-sample high-resolution vehicle track robust reconstruction method as recited in claim 5, wherein in S2-2, the upstream reference track X upREF And a downstream reference trajectory X downREF The determining process of (1) comprises:
a. if n=1, x upREF ＝Y i ，
b. If n=n,X downREF ＝Y i+1 ；
c. if N is not equal to 1 and N is not equal to N,
wherein the method comprises the steps ofIs the reconstruction track of the n-1 non-network vehicle; />Is the candidate track +.>Candidate trajectories generated for reference, +.>Is a downstream network-connected track Y i+1 Candidate trajectories generated for the reference;
s2-3 according to the upstream reference track X upREF Generating candidate tracks of the nth non-networked vehicleThe process of (1) is as follows:
a. from the following componentsCalculation of/>
b. From the following componentsCalculate->
min location error＝|x′ upREF (t-2)-x upREF (t-2)|
Wherein, initiallyIs provided by coil detection data->And->Acceleration, speed and position at time t estimated by non-networked vehicle n with reference to preceding vehicle, v upREF (t) and x upREF (t) the speed and position of the preceding vehicle of the non-networked vehicle n at the moment t, a ', v ' and x ' are estimated values, and the maximum acceleration a takes 2.75m/s 2 The most comfortable deceleration b takes 2.25m/s 2 Free flow vehicle speed v 0 Taking 32m/s, s 0 The distance between safety vehicles is 8m, s * The vehicle distance is 1.1s for the reaction time T.
7. The full-sample high-resolution vehicle track robust reconstruction method as recited in claim 5, wherein in S2-4, the reference track X is based on the downstream reference track downREF Generating candidate tracks of the nth non-networked vehicleThe process of (1) is as follows:
a. from the following componentsCalculate->
min location error＝|x′ dounREF (t+2)-x downREF (t+2)|
b. From the following componentsCalculate->
Wherein, initiallyIs provided by coil detection data->And->Acceleration, speed and position at time t, v, estimated by non-networked vehicle n with reference to rear vehicle downREF (t) and x downREF (t) speed and position of the following vehicle of the non-networked vehicle n at the time t.
8. The full-sample high-resolution vehicle track robust reconstruction method according to claim 1, wherein in S3, the specific process is:
s3-1: the space-time velocity matrix V estimated in S1 refer As constraint, solving candidate track of non-internet-connected vehicle n in S2And->Weight of +.>And->
S3-2: high-resolution track of non-internet-connected vehicle n is calculated according to weighting method
S3-3: returning to S2 to reconstruct the track of the next non-networked vehicle until the reconstruction intervalThe N non-networked vehicles in (a) are all reconstructed, namely n=n;
s3-4: let i=i+1, n=1, return S2 to continue reconstructing the trajectories of N non-networked vehicles in the next section until all non-networked vehicles in the section are reconstructed, i.e. i=i-1.
9. An electronic device comprising a memory, a processor, wherein the processor is configured to execute a program in the memory, thereby implementing the full-sample high-resolution vehicle trajectory robust reconstruction method according to any one of claims 1 to 8.
10. A storage medium containing computer executable instructions which, when executed by a computer processor, are for performing the full-sample high-resolution vehicle trajectory robust reconstruction method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311425411.0A CN117473741A (en) | 2023-10-31 | 2023-10-31 | Full-sample high-resolution vehicle track robust reconstruction method, device and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311425411.0A CN117473741A (en) | 2023-10-31 | 2023-10-31 | Full-sample high-resolution vehicle track robust reconstruction method, device and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117473741A true CN117473741A (en) | 2024-01-30 |
Family
ID=89630528
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311425411.0A Pending CN117473741A (en) | 2023-10-31 | 2023-10-31 | Full-sample high-resolution vehicle track robust reconstruction method, device and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117473741A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117334051A (en) * | 2023-10-26 | 2024-01-02 | 江苏中路交通发展有限公司 | Highway vehicle track reconstruction method and system |
-
2023
- 2023-10-31 CN CN202311425411.0A patent/CN117473741A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117334051A (en) * | 2023-10-26 | 2024-01-02 | 江苏中路交通发展有限公司 | Highway vehicle track reconstruction method and system |
| CN117334051B (en) * | 2023-10-26 | 2024-05-10 | 江苏中路交通发展有限公司 | Highway vehicle track reconstruction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ramezani et al. | Queue profile estimation in congested urban networks with probe data | |
| Feng et al. | A review and comparative study on probabilistic object detection in autonomous driving | |
| Tan et al. | Cycle-based queue length estimation for signalized intersections using sparse vehicle trajectory data | |
| Kim et al. | Extracting vehicle trajectories using unmanned aerial vehicles in congested traffic conditions | |
| Lin et al. | A Real‐Time Vehicle Counting, Speed Estimation, and Classification System Based on Virtual Detection Zone and YOLO | |
| Sun et al. | Vehicle trajectory reconstruction for signalized intersections using mobile traffic sensors | |
| Li et al. | Freeway travel-time estimation based on temporal–spatial queueing model | |
| Luo et al. | Queue length estimation for signalized intersections using license plate recognition data | |
| Hao et al. | Long queue estimation for signalized intersections using mobile data | |
| Goodall et al. | Microscopic estimation of freeway vehicle positions from the behavior of connected vehicles | |
| US20170124863A1 (en) | System and method for traffic signal timing estimation | |
| KR20200122987A (en) | Method and device for training trajectory classification model, and electronic apparatus | |
| CN117473741A (en) | Full-sample high-resolution vehicle track robust reconstruction method, device and medium | |
| CN113762473B (en) | Complex scene driving risk prediction method based on multi-time space diagram | |
| Chaudhary et al. | Video‐based road traffic monitoring and prediction using dynamic Bayesian networks | |
| Bharadwaj et al. | Traffic data analysis using image processing technique on Delhi–Gurgaon expressway | |
| Lan et al. | A new framework for traffic anomaly detection | |
| Guo et al. | Modeling driver’s evasive behavior during safety–critical lane changes: Two-dimensional time-to-collision and deep reinforcement learning | |
| CN105354542A (en) | Method for detecting abnormal video event in crowded scene | |
| Li et al. | Vehicle counting and traffic flow parameter estimation for dense traffic scenes | |
| Chang et al. | Driving safety monitoring and warning for connected and automated vehicles via edge computing | |
| Ramezani et al. | Exploiting probe data to estimate the queue profile in urban networks | |
| Dahiya et al. | Analysis of the single-regime speed-density fundamental relationships for varying spatiotemporal resolution using Zen Traffic Data | |
| CN106033643A (en) | Data processing method and data processing device | |
| Sheng et al. | Real-time queue length estimation with trajectory reconstruction using surveillance data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |