CN114783183B - Traffic situation algorithm-based monitoring method and system - Google Patents
Traffic situation algorithm-based monitoring method and system Download PDFInfo
- Publication number
- CN114783183B CN114783183B CN202210395048.1A CN202210395048A CN114783183B CN 114783183 B CN114783183 B CN 114783183B CN 202210395048 A CN202210395048 A CN 202210395048A CN 114783183 B CN114783183 B CN 114783183B
- Authority
- CN
- China
- Prior art keywords
- congestion
- traffic
- speed
- time
- situation
- 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.)
- Active
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 50
- 238000012544 monitoring process Methods 0.000 title claims abstract description 35
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000008859 change Effects 0.000 claims abstract description 36
- 230000002159 abnormal effect Effects 0.000 claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 9
- 230000007246 mechanism Effects 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims description 37
- 238000010586 diagram Methods 0.000 claims description 8
- 230000000903 blocking effect Effects 0.000 claims description 7
- 229920006395 saturated elastomer Polymers 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 description 7
- 238000010276 construction Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention provides a monitoring method and a system based on a traffic situation algorithm, wherein the method comprises the steps of calculating regional traffic situations in real time, acquiring real-time and historical traffic data, utilizing a traffic flow model to represent the change rule of regional traffic flow, calculating to obtain the characteristic value of the traffic flow model, and further estimating the regional traffic running state based on average vehicle speed, and monitoring the change of road running state in real time; and an abnormal congestion early warning step, namely establishing a short-time traffic speed prediction model based on a traffic situation algorithm by utilizing a convolutional neural network and a long-time memory network and combining an attention mechanism, performing iterative training on the established short-time traffic speed prediction model, and calculating the change situation of the traffic situation, so as to predict the speed situation of a future target road section, further judging the congestion situation of the future target road section according to the preset average speed grade, and performing early warning on the congestion situation in time, thereby facilitating the relief of traffic congestion.
Description
Technical Field
The invention relates to the technical field of intelligent highways, in particular to a traffic situation algorithm-based monitoring method and system.
Background
Under the state of rapid development of the Internet information industry, great effort is made in the aspects of research and construction of intelligent highways in China, the pace is continuously accelerated, policy system environments, resource conditions and the like of the construction of the intelligent highways are increasingly mature, and the intelligent highways enter a new era of rapid development, and the core of the intelligent highways is a cloud control platform. In recent years, with the rapid increase of the number of motor vehicles in China, highway traffic jam has become a common phenomenon, and time delay, environmental pollution and the like caused by the traffic jam can cause huge economic loss to society. Although many provinces and cities have increased the investment in the construction of traffic infrastructure, it is impossible to fundamentally solve the contradiction between the increasing traffic demand and the relatively low traffic utilization.
In the aspect of traffic situation calculation, the existing algorithm is to calculate the occupancy of adjacent coils and compare the occupancy with a calibrated threshold value so as to judge whether traffic events exist in the road section range. The calibration of the threshold value in the algorithm is particularly difficult, and particularly in a large road network, each independent threshold value is required to be set respectively according to different road geometric conditions, and the algorithm cannot represent the specific traffic mode in road sections such as an entrance ramp, an interweaving section, an ascending slope and the like, so that the false alarm rate is increased. The McMaster situation recognition algorithm also derives a flow-occupancy template consisting of four regions, each representing a particular traffic condition. The algorithm takes into account traffic flow and road geometry variations when defining blocking and non-blocking boundaries, and therefore requires redefinition of different locations and different data sets, which results in poor portability, and secondly the predefined blocking and non-blocking boundaries cannot change over time during real-time operation.
The current traffic situation of the current road section published by the prior art cannot be accurately displayed, so that the traffic situation is researched, an important basis can be provided for traveler path selection and traffic manager traffic guiding, and the method is an important application in the technical field of intelligent roads.
Disclosure of Invention
In order to solve the problems of low traffic situation real-time performance and low abnormal congestion early warning accuracy in the existing traffic situation calculation, the invention provides a traffic situation algorithm-based monitoring method. The invention also relates to a monitoring method and a system based on the traffic situation algorithm.
The technical scheme of the invention is as follows:
The traffic situation algorithm-based monitoring method is characterized by comprising the following steps of:
A real-time calculation step of regional traffic situation, namely acquiring real-time and historical traffic data, utilizing a traffic flow model to represent the change rule of regional traffic flow in the form of a scatter diagram and a graph, obtaining a functional relation among flow, speed and density in a data fitting mode, calculating to obtain a characteristic value of the traffic flow model, further evaluating the regional traffic running state based on average speed, and monitoring the change of the road running state in real time;
And an abnormal congestion early warning step, based on the calculated average speed, a short-time traffic speed prediction model based on a traffic situation algorithm is established by utilizing a convolutional neural network and a long-time memory network and combining an attention mechanism, iterative training is carried out on the established short-time traffic speed prediction model, the time and space characteristics of the actual speed of the target road section at each window moment are extracted, the hidden space association characteristics of the target road section at the window moment are mapped and output, the change condition of the traffic situation is calculated, and therefore the speed condition of the target road section in the future is predicted, the congestion condition of the target road section in the future is judged in a grading mode according to the preset average speed, and early warning is carried out on the congestion condition.
Preferably, in the step of calculating the regional traffic situation in real time, after calculating the characteristic value of the traffic flow model, the regional traffic running state is comprehensively estimated based on the average speed and by combining the calculated road network saturation and the traffic saturation.
Preferably, in the step of calculating the regional traffic situation in real time, the average vehicle speed obtained by calculation is classified according to the preset average speed, and the average vehicle speed comprises five grades including free flow, basically smooth, slight congestion, medium congestion and serious congestion, the road network saturation is the ratio of the mileage of a road section for processing the serious congestion and the medium congestion in the road network to the total mileage of the road network, and the traffic saturation is the ratio of the actual traffic flow to the saturated traffic capacity and converts different types of vehicles into standard vehicle equivalent numbers in calculation.
Preferably, in the abnormal congestion early warning step, when the change condition of traffic situation is calculated, the time period of observing the congestion in the area is counted, all the observation time period numbers are compared to obtain the congestion proportion, the area is divided into an overutilized area, a reasonable utilized area and a low utilized area, the congestion proportion difference value between the peak time period and the flat peak time period of the overutilized area exceeds a set threshold value, the congestion proportion difference value between the peak time period and the flat peak time period of the reasonable utilized area is in the set threshold value range, and the congestion proportion difference value between the peak time period and the flat peak time period of the low utilized area is lower than the set threshold value.
Preferably, in the abnormal congestion early warning step, the preset average speed is classified into five levels of free flow, basically smooth, slight congestion, medium congestion and serious congestion, and the congestion index of the future target road section is judged according to the preset average speed in a grading manner so as to obtain the congestion condition of the future target road section.
Preferably, in the step of abnormal congestion early warning, early warning is carried out on the congestion condition, and future road network operation indexes, congestion length and congestion time are also predicted; the calculation of the predicted future congestion length directly adopts the way of the summation of the congestion road section lengths.
A monitoring system based on traffic situation algorithm is characterized by comprising a regional traffic situation real-time calculation module and an abnormal congestion early warning module which are connected with each other,
The regional traffic situation real-time calculation module is used for acquiring real-time and historical traffic data, representing the change rule of regional traffic flow by using a traffic flow model in the form of a scatter diagram and a graph, obtaining a functional relation among flow, speed and density in a data fitting mode, calculating to obtain a characteristic value of the traffic flow model, further estimating the regional traffic running state based on the average speed, and monitoring the road running state change in real time;
The abnormal congestion early warning module is used for establishing a short-time traffic speed prediction model based on a traffic situation algorithm by utilizing a convolutional neural network and a long-time memory network and combining an attention mechanism based on the calculated average speed, performing iterative training on the established short-time traffic speed prediction model, extracting time and space characteristics of the actual speed of a target road section at each window moment, mapping hidden space association characteristics of the target road section at the output window moment, calculating the change situation of traffic situation, predicting the speed situation of the target road section in the future, judging the congestion situation of the target road section in the future according to the preset average speed, and carrying out early warning on the congestion situation.
Preferably, in the real-time calculation module of the regional traffic situation, the calculated average speed is classified according to preset average speed, and the average speed comprises five grades including free flow, basically smooth, slight congestion, medium congestion and serious congestion, and the regional traffic running state is comprehensively estimated based on the average speed by combining with calculation of road network saturation and traffic saturation; the road network saturation is the ratio of the mileage of a road section in which severe congestion and moderate congestion are treated in the road network to the total mileage of the road network, and the traffic saturation is the ratio of the actual traffic flow to the saturated traffic capacity and converts different types of vehicles into standard vehicle equivalent numbers in calculation.
Preferably, in the abnormal congestion early warning module, when the change condition of traffic situation is calculated, the time period of observing the congestion in the area is counted, all the observation time period numbers are compared to obtain the congestion proportion, the area is divided into an overutilized area, a reasonable utilized area and a low utilized area, the congestion proportion difference value between the peak time period and the flat peak time period of the overutilized area exceeds a set threshold value, the congestion proportion difference value between the peak time period and the flat peak time period of the reasonable utilized area is in a set threshold value range, and the congestion proportion difference value between the peak time period and the flat peak time period of the low utilized area is lower than the set threshold value.
Preferably, in the abnormal congestion early warning module, early warning is carried out on the congestion condition, and future road network operation indexes, congestion length and congestion time are also predicted; the calculation of the predicted future congestion length directly adopts the way of the summation of the congestion road section lengths.
The beneficial effects of the invention are as follows:
The invention provides a monitoring method based on a traffic situation algorithm, which utilizes a traffic flow model to represent the change rule of regional traffic flow according to real-time and duration traffic data, calculates the characteristic value of the traffic flow model, namely, calculates the characteristic value of the whole traffic flow of a road through the algorithm, evaluates the traffic running state of the road and monitors the change of the running state of the road in real time; meanwhile, according to real-time traffic information, the change situation of a future traffic situation is calculated by establishing a short-time traffic speed prediction model based on a traffic situation algorithm, the speed situation of the target road section in the future is predicted, the congestion situation of the target road section in the future is judged according to the preset average speed grade, timely early warning is carried out on the congestion situation in a future period of time, and the congestion index, the congestion length and the congestion time which are possibly caused can be further predicted, so that the traffic congestion can be relieved conveniently. The real-time traffic situation is obtained by calculation through a traffic map model according to the real-time traffic information, the real-time traffic situation calculation algorithm is carried out, and the algorithm output delay is less than 5 seconds; the real-time road condition update time is less than 2 minutes. Short-time traffic road condition prediction algorithm, wherein the algorithm output delay is less than 5 seconds; and predicting the traffic flow and the running speed in the short-time road section by taking 15min and 1h as time intervals. The system has convenience, and only needs to use the existing traffic data of the expressway: such as highway entrance and exit data, ETC portal data, internet traffic data, etc., without adding new equipment. The method comprises the steps of predicting the future congestion condition of a certain road section, predicting the vehicle speed condition of the road section in the future, judging the congestion condition of a circuit network according to the vehicle speed, and calculating the change condition of the traffic situation according to the conventional model of time sequence feature extraction, wherein the prediction of the future vehicle speed is completed by adopting a short-time traffic speed prediction model GCN-LSTM based on a traffic situation algorithm, combining a convolutional neural network GCN with a long-short-term memory network LSTM, modeling an expressway network, extracting the space features of the actual speed of a target road section at each window moment by utilizing the GCN neural network, mapping and outputting the hidden space associated features of the target road section at the window moment, wherein the time dimension has time correlation of the speed, and the LSTM is used as a conventional model of time sequence feature extraction and is used as a key method of time feature extraction to calculate the change condition of the traffic situation, so that the vehicle speed condition of the target road section in the future is predicted, and judging the congestion condition of the target road section in the future according to the preset average speed classification. The method has the advantages that the accuracy is realized, and the real-time road condition and abnormal congestion early warning accuracy rate reaches 80%; the road condition prediction accuracy rate of the road within 15 minutes is 80%; the accuracy rate of predicting the road congestion within one hour reaches 90 percent. According to the invention, the road information and the multidimensional internet data are fused, the discovery, recognition, understanding and analysis, response and disposal capabilities of traffic jams are improved from the global view, and meanwhile, the results calculated by the algorithm can be displayed to the manager in real time, so that the manager can make decisions in time.
The invention also relates to a monitoring system based on the traffic situation algorithm, which corresponds to the monitoring method based on the traffic situation algorithm, and can be understood as a system for realizing the monitoring method based on the traffic situation algorithm, comprising a regional traffic situation real-time calculation module and an abnormal congestion early warning module, wherein the modules work cooperatively with each other, the regional traffic situation real-time calculation is carried out through a traffic flow model, the road running state change is monitored in real time, the future speed is predicted by combining a short-time traffic speed prediction model GCN-LSTM based on the traffic situation algorithm, the future congestion condition of a certain road section can be predicted, the traffic situation algorithm can obtain detailed traffic running condition data, the real-time monitoring of the traffic condition by a road manager is facilitated, the predicted road section future traffic condition can provide basis for the path selection of a traveler, the traffic congestion is relieved, the prediction accuracy is improved, and the system is convenient to use.
Drawings
Fig. 1 is a flow chart of a traffic situation algorithm-based monitoring method of the present invention.
Fig. 2 is a preferred flow chart of the traffic situation algorithm based monitoring method of the present invention.
FIG. 3 is a flowchart of the short-term traffic speed prediction model GCN-LSTM of the present invention.
Detailed Description
The present invention will be described below with reference to the accompanying drawings.
The invention relates to a monitoring method based on a traffic situation algorithm, and a flow chart of the method is shown in fig. 1, and the method comprises the following steps:
And a regional traffic situation real-time calculation step, namely dividing the expressway network into a plurality of grid areas, acquiring real-time and historical traffic data, utilizing a traffic flow model to represent the change rule of regional traffic flow in the form of a scatter diagram and a graph, obtaining a functional relation among flow, speed and density in a data fitting mode, calculating to obtain a characteristic value of the traffic flow model, further evaluating the regional traffic running state based on the average speed, and monitoring the change of the road running state in real time. The traffic data mainly comprises an expressway outlet flow water meter, an expressway inlet flow water meter, ETC portal identification data, internet dynamic road condition data, traffic event data, traffic jam early warning, weather data and the like.
In the running process of the highway traffic flow, the traffic state is always changed, and the parameters capable of representing the traffic state are various, including flow, speed, density, occupancy, time interval, travel time and the like. The invention calculates the regional traffic flow based on a statistical traffic flow basic diagram model (i.e. a traffic flow model, a Van-Aerde model), describes the relationship among three traffic elements of flow, speed and density, characterizes the state and change rule of the traffic flow in the form of a scatter diagram and a graph, and obtains the functional relationship among the flow, the speed and the density in a data fitting mode.
The invention collects the speed and the passing direction of the vehicle in real time and describes the regional traffic flow by utilizing a Van-Aerde model, and the model formula is as follows:
Wherein: s represents a vehicle head distance (km); v denotes an operation speed (km/h); v f denotes the free stream speed (km/h); c 1 denotes a headstock distance parameter (km); c 2 denotes a variable following parameter (km 2/h);c3 denotes a headway parameter (h -1).
The calculation method of the parameter c 1、c2、c3 is as follows:
c1＝vf(2vm-vf)/(kjvm 2) (2)
c2＝vf(vf-vm)/(kjvm 2) (3)
Wherein: q m is the maximum flow rate (veh/h); v m is the speed (km/h) corresponding to the maximum flow rate; v f is the free flow speed (km/h); k j is the blocking density (veh/km).
Density k is the inverse of the head spacing:
the flow q can be derived from the speed, density:
q＝kv (6)
the model parameters are calibrated by collecting expressway vehicle data in real time to obtain functions, and then a flow rate-speed curve, a flow rate-density curve and a speed-density curve are obtained.
After the overall traffic flow characteristics of the area are obtained, the running state of the traffic flow of the area is evaluated, and whether the traffic flow runs smoothly and whether traffic jams are generated or not is analyzed. The average speed is an index which can represent the running state of the vehicle most, the average speed is used for selecting the average travel speed of the road network, and the average travel speed refers to the ratio of the total travel mileage of all vehicles in the road or the road network to the total travel time, wherein the travel time comprises the intermediate parking time and the queuing time. The speed is high, the running of the traffic flow is free, and the traffic flow is not influenced by the environment and road traffic conditions; the lower speed indicates that the traffic flow is obstructed from running and cannot be maintained at a higher speed.
Traffic conditions, in other words, congestion conditions, are classified into 5 classes according to average speed according to different speed limit (km/h), and the average speed classification table is shown in table 1.
TABLE 1
The average speed adopts the average travel speed of the road section, is calculated mainly according to the highway exit flow chart, and meanwhile, the vehicle types are distinguished, and bad value data are removed. The main flow is as follows:
(1) Selecting data of a corresponding interval and a corresponding time period (real time);
(2) Matching the running road section length s i of each vehicle according to the serial number of the toll station at the vehicle entrance and the toll station, and calculating the running time t i = outbound time-inbound time of each vehicle;
(3) Setting the driving time t i to be 0-24 h, and eliminating data exceeding the driving time t i;
(4) According to Calculating the average speed of each vehicle, setting the range of the average speed to be 0-150 km/h, and eliminating data exceeding the range;
(5) According to And calculating the average speed corresponding to each road section. Where n is the number of vehicles in the period of the road segment.
Further, as shown in fig. 2, after the characteristic value of the traffic flow model is calculated, the regional traffic running state is comprehensively estimated based on the average speed and by combining the calculated road network saturation and the traffic saturation. The traffic flow is calculated mainly according to ETC portal plate identification data, and meanwhile, the vehicle type and the uplink and downlink are distinguished, and the license plate numbers in a certain period are directly summed. The calculated average speed is graded according to the preset average speed, and the average speed comprises five grades of free flow, basically smooth, slightly congested, moderately congested and severely congested.
Road network saturation is also called road section load degree and road network capacity adaptability, and is the ratio of the road section mileage number in the state of serious congestion and medium congestion in the road network to the total mileage of the road network, and the calculation formula is as follows:
Wherein M is road network saturation; p is the mileage (km) of a road section in a severe congestion state and a moderate congestion state in the road network; x is the total mileage (km) of the road network.
Traffic saturation is the ratio of actual traffic flow to saturated traffic capacity and converts different types of vehicles into standard vehicle equivalent numbers, namely VC ratio, when calculated, the calculation formula is as follows:
Wherein V is the actual vehicle flow (pcu); c is the saturated traffic capacity (pcu), which is related to the design speed and number of lanes of the highway, c=single lane traffic capacity, which refers to the related design specifications, as shown in table 2.
TABLE 2
| Design speed (km/h) | Traffic capacity (pcu/h/ln) |
| 120 | 2200 |
| 100 | 2100 |
| 80 | 2000 |
| 60 | 1800 |
In the above calculations, it was necessary to convert different types of vehicles to standard vehicle equivalent numbers (pcu), as shown in table 3 with reference to the relevant standards.
TABLE 3 Table 3
Specifically, the vehicle models are classified as follows: 01-one type passenger car 2-two type passenger car 3-three type passenger car 4-four type passenger car 11-one type truck 12-two type truck 13-three type truck 14-four type truck 15-five type truck 16-six type truck 21-one type special working vehicle 22-two type special working vehicle 23-three type special working vehicle 24-four type special working vehicle 25-five type special working vehicle 26-six type special working vehicle. The conversion coefficient corresponding to each vehicle model is set as shown in table 4.
TABLE 4 Table 4
| Vehicle model | Conversion coefficient | VEHICLECLASS numbering |
| Small passenger/truck | 1.0 | 0，1，2，11，21 |
| Bus/medium truck | 1.5 | 3，4，12，13，22，23 |
| Large truck | 2.0 | 14，24 |
| Oversized truck | 3.0 | 15，16，25，26 |
And an abnormal congestion early warning step, based on the calculated average speed, a short-time traffic speed prediction model based on a traffic situation algorithm is established by utilizing a convolutional neural network and a long-time memory network and combining an attention mechanism, iterative training is carried out on the established short-time traffic speed prediction model, the time and space characteristics of the actual speed of the target road section at each window moment are extracted, the hidden space association characteristics of the target road section at the window moment are mapped and output, the change condition of the traffic situation is calculated, and therefore the speed condition of the target road section in the future is predicted, the congestion condition of the target road section in the future is judged in a grading mode according to the preset average speed, and early warning is carried out on the congestion condition.
Predicting the future congestion condition of a certain road section firstly requires predicting the vehicle speed condition of the road section in the future, and judging the congestion condition of the circuit breaker network according to the vehicle speed. The prediction of the future vehicle speed is completed by adopting a short-time traffic speed prediction model GCN-LSTM based on a graph convolution neural network, a flow chart is shown in figure 3, after preprocessing real-time traffic data and historical traffic data, test data are added for model training and testing, and the real-time traffic speed prediction is obtained. The overall framework of the model is mainly divided into 5 modules: the system comprises an input layer, a space correlation feature extraction module, a time correlation feature extraction module, an attention mechanism module and a prediction module. The model utilizes a GCN convolutional neural network to extract the spatial characteristics of the actual speed of the target road section at each window moment, and maps and outputs the hidden spatial association characteristics of the target road section at the window moment. The speed has time correlation in the time dimension, and the LSTM long-short time memory network is used as a common model for time sequence feature extraction and is used as a key method for time feature extraction.
The parameters of the GCN-LSTM model mainly comprise: the parameter configuration with highest prediction precision is obtained through continuous parameter debugging, wherein the parameter configuration comprises batch size, learning rate, iteration times, forgetting offset and LSTM unit number. For the model input layer, the training set (80%) was used as input to the training process and the remaining data (20%) was used as the test set. The GCN-LSTM model may be trained using an Adam optimizer. In the training process, the final objective is to minimize the difference between the actual traffic speed and the predicted speed on the road, using V t andThe loss functions of the GCN-LSTM model, representing the actual and predicted speeds, respectively, are shown below:
Where L reg is an L2 regularization term that prevents overfitting and λ is a hyper-parameter.
The evaluation index is to evaluate the predicted performance of the model, and the actual speed (V t) and the predicted speed are evaluated using the following three indexesErrors between them.
(1) Root Mean Square Error (RMSE)
(2) Mean Absolute Error (MAE)
(3) Determining coefficient (R 2)
Specifically, RMSE and MAE are used to measure prediction error, the smaller the value, the better, and the smaller the representative error; while the larger R 2 is, the better, when the predictive model does not make any errors, the value of R 2 may be 1, but is often not present in reality.
The prediction of the running speed is performed by subdividing the road section, and then the congestion index of the road section in the future is judged by table 1 (average speed classification table), and the road section is equally classified into 5 classes: free flow, substantially clear, lightly congested, moderately congested, severely congested. For the possible congestion situation in the future, the method adopts the method of predicting the speed of the road section in the future, judging the congestion situation according to the speed, and directly adopting the method of adding and summing the congestion road sections for calculating the congestion length. The traffic situation algorithm can obtain detailed traffic running condition data, so that a road manager can conveniently monitor traffic conditions in real time, and predicted future traffic conditions of road sections can provide basis for route selection of travelers and relieve traffic jams.
According to the traffic real-time traffic information, calculating the change condition of traffic situation, early warning the congestion condition in time, and predicting the road network operation index, the congestion length and the congestion time. Each region is described in units of 2 minutes in time, and the change trend of the region operation state in time is obtained. And researching the rule of congestion in each area, and analyzing the frequent and sporadic congestion to obtain the regional traffic property. Counting the time period when the area observes the congestion, comparing all the observed time periods to obtain the congestion proportion, and dividing the area into three types as shown in table 5.
TABLE 5
| Region type | Description of: |
| Overutilized region | The congestion ratio difference between the peak time and the flat peak time is larger |
| Reasonable utilization area | The congestion ratio difference between the peak time and the flat peak time is in a reasonable range |
| Low utilization region | The congestion ratio in peak time is hardly increased compared with Yu Pingfeng time |
The invention also relates to a monitoring system based on the traffic situation algorithm, which corresponds to the monitoring method based on the traffic situation algorithm, and can be understood as a system for realizing the monitoring method based on the traffic situation algorithm, comprising a regional traffic situation real-time calculation module and an abnormal congestion early warning module,
The regional traffic situation real-time calculation module is used for acquiring real-time and historical traffic data, representing the change rule of regional traffic flow by using a traffic flow model in the form of a scatter diagram and a graph, obtaining a functional relation among flow, speed and density in a data fitting mode, calculating to obtain a characteristic value of the traffic flow model, further estimating the regional traffic running state based on the average speed, and monitoring the road running state change in real time. Preferably, the calculated average speed is graded according to preset average speed, and the average speed comprises five grades of free flow, basically smooth, slight congestion, medium congestion and serious congestion, and the regional traffic running state is comprehensively evaluated based on the average speed and combined with the calculated road network saturation and traffic saturation; the road network saturation is the ratio of the mileage of a road section in which severe congestion and moderate congestion are treated in the road network to the total mileage of the road network, and the traffic saturation is the ratio of the actual traffic flow to the saturated traffic capacity and converts different types of vehicles into standard vehicle equivalent numbers in calculation.
The abnormal congestion early warning module is used for establishing a short-time traffic speed prediction model based on a traffic situation algorithm by utilizing a convolutional neural network and a long-time memory network and combining an attention mechanism based on the calculated average speed, performing iterative training on the established short-time traffic speed prediction model, extracting time and space characteristics of the actual speed of a target road section at each window moment, mapping hidden space association characteristics of the target road section at the output window moment, calculating the change situation of traffic situation, predicting the speed situation of the target road section in the future, judging the congestion situation of the target road section in the future according to the preset average speed, and carrying out early warning on the congestion situation.
Further, in the abnormal congestion early warning module, when the change condition of traffic situation is calculated, the time interval of observing the congestion in the area is counted, all the observation time intervals are compared to obtain the congestion proportion, the area is divided into an overutilized area, a reasonable utilized area and a low utilized area, the congestion proportion difference value between the peak time interval and the flat peak time interval of the overutilized area exceeds a set threshold value, the congestion proportion difference value between the peak time interval and the flat peak time interval of the reasonable utilized area is in the set threshold value range, and the congestion proportion difference value between the peak time interval and the flat peak time interval of the low utilized area is lower than the set threshold value.
Further, in the abnormal congestion early warning module, the preset average speed is divided into five levels of free flow, basically smooth, slight congestion, medium congestion and serious congestion, and the congestion index of the future target road section is judged in a grading mode according to the preset average speed so as to obtain the congestion condition of the future target road section. The running index, the congestion length and the congestion time of the road network can be predicted; the calculation of the future congestion length directly adopts a mode of summing the length of the congestion road sections.
It should be noted that the above-described embodiments will enable those skilled in the art to more fully understand the invention, but do not limit it in any way. Therefore, although the present invention has been described in detail with reference to the drawings and examples, it will be understood by those skilled in the art that the present invention may be modified or equivalent, and in all cases, all technical solutions and modifications which do not depart from the spirit and scope of the present invention are intended to be included in the scope of the present invention.
Claims (9)
1. The traffic situation algorithm-based monitoring method is characterized by comprising the following steps of:
Real-time calculation of regional traffic situation, namely acquiring real-time and historical traffic data, utilizing a traffic flow model to represent the change rule of regional traffic flow in the form of a scatter diagram and a graph, obtaining a functional relation among flow, speed and density in a data fitting mode, calculating to obtain a characteristic value of the traffic flow model, further comprehensively evaluating the regional traffic running state based on average speed and combining calculation of road network saturation and traffic saturation, and monitoring the change of the road running state in real time, wherein the road network saturation is the ratio of the mileage of a road section with severe congestion and medium congestion to the total mileage of the road network, and the traffic saturation is the ratio of actual traffic flow to saturated traffic capacity and converts different types of vehicles into standard vehicle equivalent numbers according to respective corresponding conversion coefficients during calculation; the traffic flow model formula is as follows:
Wherein: s represents the distance between the vehicle heads; v represents the running speed; v f denotes the free flow velocity; c 1 represents a vehicle head space parameter; c 2 denotes a variable following parameter; c 3 represents a headway parameter;
the calculation method of the parameter c 1、c2、c3 is as follows:
c1＝vf(2vm-vf)/(kjvm 2)
c2＝vf(vf-vm)/(kjvm 2)
Wherein: q m is the maximum flow rate; v m is the corresponding speed at which the flow rate is maximum; v f is the free flow velocity; k j is the blocking density;
An abnormal congestion early warning step, based on the calculated average speed, a short-time traffic speed prediction model based on a traffic situation algorithm is established by utilizing a convolutional neural network and a long-time memory network and combining an attention mechanism, and parameters of the short-time traffic speed prediction model comprise: batch size, learning rate, iteration times, forgetting offset and LSTM unit number, extracting time and space characteristics of actual speed of a target road section at each window moment through continuous parameter debugging and iterative training on an established short-time traffic speed prediction model, mapping hidden space association characteristics of the target road section at the output window moment, calculating the change condition of traffic situation, predicting the speed condition of the target road section in the future, judging the congestion condition of the target road section in the future according to preset average speed in a grading mode, and early warning the congestion condition.
2. The traffic situation algorithm-based monitoring method according to claim 1, wherein in the regional traffic situation real-time calculation step, the calculated average vehicle speed is further classified according to a preset average speed, and the five classes include free flow, basic smoothness, slight congestion, medium congestion and severe congestion.
3. The traffic situation algorithm-based monitoring method according to claim 1 or 2, wherein in the abnormal congestion pre-warning step, when the change situation of the traffic situation is calculated, the time period of the observed congestion of the area is counted, the number of all the observed time periods is compared to obtain the congestion ratio, the area is divided into three types of an overutilized area, a reasonable utilized area and a low utilized area, the congestion ratio difference value between the peak time period and the flat peak time period of the overutilized area exceeds a set threshold, the congestion ratio difference value between the peak time period and the flat peak time period of the reasonable utilized area is within the set threshold, and the congestion ratio difference value between the peak time period and the flat peak time period of the low utilized area is lower than the set threshold.
4. The traffic situation algorithm-based monitoring method according to claim 1, wherein in the abnormal congestion pre-warning step, the preset average speed is divided into five levels of free flow, basically smooth, slight congestion, medium congestion and serious congestion, and the congestion index of the future target road section is judged in a grading manner according to the preset average speed so as to obtain the congestion condition of the future target road section.
5. The traffic situation algorithm-based monitoring method according to claim 4, wherein in the abnormal congestion pre-warning step, pre-warning is performed on the congestion condition, and future road network operation indexes, congestion lengths and congestion time are predicted; the calculation of the predicted future congestion length directly adopts the way of the summation of the congestion road section lengths.
6. A monitoring system based on traffic situation algorithm is characterized by comprising a regional traffic situation real-time calculation module and an abnormal congestion early warning module which are connected with each other,
The regional traffic situation real-time calculation module is used for acquiring real-time and historical traffic data, representing the change rule of regional traffic flow by using a traffic flow model in the form of a scatter diagram and a graph, obtaining a functional relation among flow, speed and density in a data fitting mode, calculating to obtain a characteristic value of the traffic flow model, further comprehensively evaluating the regional traffic running state based on average speed and combining with calculation of road network saturation and traffic saturation, and monitoring the change of the road running state in real time, wherein the road network saturation is the ratio of the mileage of a road section with severe congestion and medium congestion in a road network to the total mileage of the road network, and the traffic saturation is the ratio of actual traffic flow to saturated traffic capacity and converts different types of vehicles into standard vehicle equivalent numbers according to corresponding conversion coefficients in calculation; the traffic flow model formula is as follows:
Wherein: s represents the distance between the vehicle heads; v represents the running speed; v f denotes the free flow velocity; c 1 represents a vehicle head space parameter; c 2 denotes a variable following parameter; c 3 represents a headway parameter;
the calculation method of the parameter c 1、c2、c3 is as follows:
c1＝vf(2vm-vf)/(kjvm 2)
c2＝vf(vf-vm)/(kjvm 2)
Wherein: q m is the maximum flow rate; v m is the corresponding speed at which the flow rate is maximum; v f is the free flow velocity; k j is the blocking density;
The abnormal congestion early warning module is used for establishing a short-time traffic speed prediction model based on a traffic situation algorithm by utilizing a convolutional neural network and a long-time memory network and combining an attention mechanism based on the calculated average speed, and parameters of the short-time traffic speed prediction model comprise: batch size, learning rate, iteration times, forgetting offset and LSTM unit number, extracting time and space characteristics of actual speed of a target road section at each window moment through continuous parameter debugging and iterative training on an established short-time traffic speed prediction model, mapping hidden space association characteristics of the target road section at the output window moment, calculating the change condition of traffic situation, predicting the speed condition of the target road section in the future, judging the congestion condition of the target road section in the future according to preset average speed in a grading mode, and early warning the congestion condition.
7. The traffic situation algorithm-based monitoring system according to claim 6, wherein in the regional traffic situation real-time calculation module, the calculated average vehicle speed is further classified according to a preset average speed, and the five classes include free flow, basic smoothness, slight congestion, medium congestion and severe congestion.
8. The traffic situation algorithm-based monitoring system according to claim 6 or 7, wherein in the abnormal congestion early warning module, when the change situation of the traffic situation is calculated, the time period of the observed congestion of the area is counted, the number of all the observed time periods is compared to obtain the congestion ratio, the area is divided into three types of an overutilized area, a reasonable utilized area and a low utilized area, the congestion ratio difference value between the peak time period and the flat peak time period of the overutilized area exceeds a set threshold, the congestion ratio difference value between the peak time period and the flat peak time period of the reasonable utilized area is within the set threshold, and the congestion ratio difference value between the peak time period and the flat peak time period of the low utilized area is lower than the set threshold.
9. The traffic situation algorithm-based monitoring system according to claim 8, wherein in the abnormal congestion pre-warning module, pre-warning is performed on congestion conditions, and future road network operation indexes, congestion lengths and congestion times are predicted; the calculation of the predicted future congestion length directly adopts the way of the summation of the congestion road section lengths.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210395048.1A CN114783183B (en) | 2022-04-15 | 2022-04-15 | Traffic situation algorithm-based monitoring method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210395048.1A CN114783183B (en) | 2022-04-15 | 2022-04-15 | Traffic situation algorithm-based monitoring method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114783183A CN114783183A (en) | 2022-07-22 |
| CN114783183B true CN114783183B (en) | 2024-05-24 |
Family
ID=82429538
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210395048.1A Active CN114783183B (en) | 2022-04-15 | 2022-04-15 | Traffic situation algorithm-based monitoring method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114783183B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115394083B (en) * | 2022-08-29 | 2023-06-06 | 合肥工业大学 | Highway traffic jam prediction method based on deep learning |
| CN115410369B (en) * | 2022-08-30 | 2024-02-20 | 合肥工业大学 | Real-time road traffic speed interval construction method |
| CN116013083B (en) * | 2023-01-10 | 2024-03-12 | 合肥工业大学 | Road network frequent traffic state identification method based on electronic map data |
| CN115830874B (en) * | 2023-02-10 | 2023-05-12 | 西南交通大学 | Traffic flow basic diagram fitting performance evaluation method and system |
| CN116631196B (en) * | 2023-07-25 | 2023-11-28 | 南京农业大学 | Traffic road condition prediction method and device based on big data |
| CN117196153A (en) * | 2023-09-25 | 2023-12-08 | 广东汇通信息科技股份有限公司 | Intelligent traffic comprehensive management system based on traffic big data |
| CN117292551B (en) * | 2023-11-27 | 2024-02-23 | 辽宁邮电规划设计院有限公司 | Urban traffic situation adjustment system and method based on Internet of things |
| CN117831287B (en) * | 2023-12-29 | 2024-05-31 | 北京大唐高鸿数据网络技术有限公司 | Method, device, equipment and storage medium for determining highway congestion index |
| CN117533356B (en) * | 2024-01-09 | 2024-03-29 | 北京市北太机电设备工贸有限公司 | Intelligent driving assistance system and method |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0714093A (en) * | 1993-04-28 | 1995-01-17 | Hitachi Ltd | Traffic signal controlling method |
| CN105023433A (en) * | 2015-07-01 | 2015-11-04 | 重庆大学 | Method for predicting range influenced by abnormal traffic event of highway |
| WO2017166474A1 (en) * | 2016-03-29 | 2017-10-05 | 中国科学院深圳先进技术研究院 | Method and system for intersection group-based traffic control |
| CN108305468A (en) * | 2017-01-13 | 2018-07-20 | 普天信息技术有限公司 | One kind being based on shifty traffic control method and system |
| CN110648527A (en) * | 2019-08-20 | 2020-01-03 | 浙江工业大学 | Traffic speed prediction method based on deep learning model |
| CN111540199A (en) * | 2020-04-21 | 2020-08-14 | 浙江省交通规划设计研究院有限公司 | High-speed traffic flow prediction method based on multi-mode fusion and graph attention machine mechanism |
| CN112435462A (en) * | 2020-10-16 | 2021-03-02 | 同盾控股有限公司 | Method, system, electronic device and storage medium for short-time traffic flow prediction |
| CN112820105A (en) * | 2020-12-31 | 2021-05-18 | 银江股份有限公司 | Road network abnormal area processing method and system |
| CN113643539A (en) * | 2021-08-12 | 2021-11-12 | 合肥永信科翔智能技术有限公司 | Intelligent traffic control system based on big data |
| CN113704956A (en) * | 2021-06-15 | 2021-11-26 | 深圳市综合交通设计研究院有限公司 | Urban road online microscopic simulation method and system based on digital twin technology |
| CN113947905A (en) * | 2021-10-19 | 2022-01-18 | 交通运输部公路科学研究所 | Traffic operation situation sensing method, module and system |
-
2022
- 2022-04-15 CN CN202210395048.1A patent/CN114783183B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0714093A (en) * | 1993-04-28 | 1995-01-17 | Hitachi Ltd | Traffic signal controlling method |
| CN105023433A (en) * | 2015-07-01 | 2015-11-04 | 重庆大学 | Method for predicting range influenced by abnormal traffic event of highway |
| WO2017166474A1 (en) * | 2016-03-29 | 2017-10-05 | 中国科学院深圳先进技术研究院 | Method and system for intersection group-based traffic control |
| CN108305468A (en) * | 2017-01-13 | 2018-07-20 | 普天信息技术有限公司 | One kind being based on shifty traffic control method and system |
| CN110648527A (en) * | 2019-08-20 | 2020-01-03 | 浙江工业大学 | Traffic speed prediction method based on deep learning model |
| CN111540199A (en) * | 2020-04-21 | 2020-08-14 | 浙江省交通规划设计研究院有限公司 | High-speed traffic flow prediction method based on multi-mode fusion and graph attention machine mechanism |
| CN112435462A (en) * | 2020-10-16 | 2021-03-02 | 同盾控股有限公司 | Method, system, electronic device and storage medium for short-time traffic flow prediction |
| CN112820105A (en) * | 2020-12-31 | 2021-05-18 | 银江股份有限公司 | Road network abnormal area processing method and system |
| CN113704956A (en) * | 2021-06-15 | 2021-11-26 | 深圳市综合交通设计研究院有限公司 | Urban road online microscopic simulation method and system based on digital twin technology |
| CN113643539A (en) * | 2021-08-12 | 2021-11-12 | 合肥永信科翔智能技术有限公司 | Intelligent traffic control system based on big data |
| CN113947905A (en) * | 2021-10-19 | 2022-01-18 | 交通运输部公路科学研究所 | Traffic operation situation sensing method, module and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114783183A (en) | 2022-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114783183B (en) | Traffic situation algorithm-based monitoring method and system | |
| CN109544932B (en) | Urban road network flow estimation method based on fusion of taxi GPS data and gate data | |
| WO2020244288A1 (en) | Method and apparatus for evaluating truck driving behaviour based on gps trajectory data | |
| CN103839409B (en) | Based on the traffic flow modes method of discrimination of multibreak facial vision sensing cluster analysis | |
| Kim et al. | Urban traffic flow prediction system using a multifactor pattern recognition model | |
| CN107085943B (en) | Short-term prediction method and system for road travel time | |
| CN104778837A (en) | Multi-time scale forecasting method for road traffic running situation | |
| CN109002622B (en) | Method for estimating overall load response of large-span bridge under random traffic flow effect | |
| CN111583628B (en) | Road network heavy truck traffic flow prediction method based on data quality control | |
| CN108648445B (en) | Dynamic traffic situation prediction method based on traffic big data | |
| CN113177657B (en) | Rail transit passenger flow prediction method and device | |
| CN113436433B (en) | Efficient urban traffic outlier detection method | |
| CN113496314B (en) | Method for predicting road traffic flow by neural network model | |
| CN113436432A (en) | Method for predicting short-term traffic risk of road section by using road side observation data | |
| CN114187766B (en) | Road service level evaluation method based on saturation rate | |
| CN106023592A (en) | Traffic jam detection method based on GPS data | |
| CN114596700A (en) | Real-time highway section flow estimation method based on portal data | |
| CN116597642A (en) | Traffic jam condition prediction method and system | |
| CN110867075B (en) | Method for evaluating influence of road speed meter on reaction behavior of driver under rainy condition | |
| CN117746626A (en) | Intelligent traffic management method and system based on traffic flow | |
| CN112183597B (en) | Unbalanced traffic accident severity analysis method based on spectral clustering | |
| CN113642162B (en) | Urban road traffic emergency plan simulation comprehensive analysis method | |
| CN110956808B (en) | Heavy truck traffic flow prediction method based on non-full-sample positioning data | |
| CN110120154B (en) | Traffic road condition prediction method using detector data under large-scale road network | |
| CN112349098A (en) | Method for estimating accident severity by environmental elements in exit ramp area of expressway |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |